Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Microbiological Risk Assessment of Raw Cow Milk Risk Assessment Microbiology Section December 2009 MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK ii TABLE OF CONTENTS ACKNOWLEDGEMENTS .................................................................................................... V ABBREVIATIONS ................................................................................................................ VI EXECUTIVE SUMMARY ...................................................................................................... 1 1. BACKGROUND................................................................................................................... 4 2. INTRODUCTION ................................................................................................................ 5 2.1 2.2 2.3 2.4 PURPOSE...................................................................................................................... 5 SCOPE.......................................................................................................................... 5 DEFINITION OF RAW MILK ........................................................................................... 5 APPROACH .................................................................................................................. 6 3. FINDINGS ............................................................................................................................ 7 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 PRODUCTION OF RAW MILK ......................................................................................... 7 CONSUMPTION OF RAW MILK....................................................................................... 8 CONSUMPTION OF PASTEURISED MILK ......................................................................... 8 MICROBIOLOGICAL HAZARDS IN RAW COW MILK ........................................................ 9 PREVALENCE OF PATHOGENS IN RAW COW MILK ....................................................... 11 MILK PRODUCTION AND ITS IMPACT ON MILK SAFETY ............................................... 15 POST-MILKING FACTORS IMPACTING ON MILK SAFETY .............................................. 17 OTHER FACTORS IMPACTING ON MILK SAFETY .......................................................... 17 4. QUANTITATIVE RISK ASSESSMENT ........................................................................ 19 4.1 4.2 4.3 4.4 4.5 4.6 4.7 INTRODUCTION .......................................................................................................... 19 MODEL OVERVIEW .................................................................................................... 19 MODEL INPUTS .......................................................................................................... 21 SIMULATION RESULTS ............................................................................................... 32 SENSITIVITY ANALYSIS FOR ON-FARM FACTORS ........................................................ 35 MEASURES TO REDUCE LEVELS OF PATHOGENS IN RAW MILK.................................... 36 VARIABILITY AND UNCERTAINTY .............................................................................. 38 5. DISCUSSION ..................................................................................................................... 39 6. DATA GAPS AND AREAS FOR FURTHER RESEARCH ......................................... 41 7. CONCLUSION ................................................................................................................... 42 ANNEX 1: SUMMARY OF MILK CONSUMPTION DATA ...................................... 44 ANNEX 2: SUMMARY OF FOODBORNE ILLNESS ASSOCIATED WITH CONSUMPTION OF RAW COW MILK ................................................... 46 ANNEX 3: SUMMARY OF KEY MICROBIOLOGICAL HAZARDS ASSOCIATED WITH RAW COW MILK ............................................................................. 49 ANNEX 4: HAZARD IDENTIFICATION AND CHARACTERISATION OF MODELLED PATHOGENS ........................................................................ 60 1. 2. 3. 4. CAMPYLOBACTER SPP. .................................................................................................... 60 ESCHERICHIA COLI (PATHOGENIC) ................................................................................. 65 LISTERIA MONOCYTOGENES ............................................................................................ 75 SALMONELLA SPP. ........................................................................................................... 81 MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK iii ANNEX 5: BULK MILK TANK PREVALENCE ......................................................... 88 ANNEX 6: WITH-IN HERD PREVALENCE ................................................................ 92 ANNEX 7: GROWTH AND INACTIVATION RATES EQUATIONS....................... 95 ANNEX 8: RISK MANAGEMENT QUESTIONS......................................................... 98 ANNEX 9: REFERENCES ............................................................................................... 99 MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK iv Acknowledgements Food Standards Australia New Zealand gratefully acknowledges the support and assistance of many contributors during the preparation of this risk assessment. Special thanks are extended to members of the Dairy Scientific Advisory Panel for their commitment, guidance and advice throughout the risk assessment process, and to Australian dairy industry personnel who graciously authorised access to industry information and State authorities for the provision of microbiological data. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK v Abbreviations CDC Centers for Disease Control and Prevention Codex Codex Alimentarius Commission EHEC Enterohaemorrhagic Escherichia coli FAO Food and Agriculture Organization of the United Nations FDA United States Food and Drug Administration FSANZ Food Standards Australia New Zealand HUS Haemolytic ureamic syndrome LP Lactoperoxidase MMWR Morbidity and Mortality Weekly Report NNS National Nutrition Survey OIE Organisation Mondiale de la Santé Animale RMSSS Roy Morgan Single Source Survey SLT Shiga-like toxin STEC Shiga toxin-producing Escherichia coli SLTEC Shiga-like toxin-producing Escherichia coli ST Shiga toxin The Profile A Risk Profile of Dairy Products in Australia VT Verotoxins VTEC Verocytotoxin-producing Escherichia coli WA Western Australia WHO World Health Organization MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK vi Executive summary The risk assessment of raw cow milk brings together information on the public health risks associated with the consumption of raw cow milk, and estimates the resulting burden of illness that may occur under current Australian production and marketing conditions. The assessment was undertaken to answer the following questions: (1) What are the risks to public health and safety posed by the consumption, in Australia, of raw cow milk? (2) What are the factors that would have the greatest impact on public health and safety along the production chain? The risk assessment considered domestic and international information from published and unpublished sources on: milk production systems, prevalence and levels of pathogens in raw cow milk and in cattle, consumption data, and epidemiological data. This information provided an overall picture of the public health risks associated with consumption of raw cow milk. In order to estimate the likelihood of illness for Australian consumers following consumption of raw cow milk, quantitative microbiological modelling was undertaken. The modelling predicted the number of illnesses per 100,000 servings of raw milk consumed directly from the bulk milk tank, after farm-gate sale, and retail sale for four identified key milkborne pathogens: Campylobacter spp., enterohaemorrhagic Escherichia coli, Salmonella spp. and Listeria monocytogenes. The key findings of the risk assessment can be summarised as: • Raw cow milk is associated with foodborne illness internationally, and has been linked to illnesses in Australia • Four key pathogens are associated with outbreaks of foodborne illness implicating raw cow milk, these are Campylobacter spp., Salmonella spp., Listeria monocytogenes and pathogenic Escherichia coli • Internationally, raw cow milk has been found to be a significant source of pathogenic microorganisms • Unpublished research suggests consumption of raw milk is likely to be low among the general population, however, certain groups preferentially consume raw milk • The burden of illness after retail purchase was predicted to be less than 1 case of campylobacteriosis, 97 cases of EHEC, 153 cases of salmonellosis and up to 170 cases of listeriosis (in a susceptible sub-population). The estimated number of cases are per 100,000 daily serves of a mean daily intake of 540 ml of milk to a child • The predicted number of illnesses following consumption of raw milk from the bulk milk tank or after farm gate sales per 100,000 daily serves is: o 19 cases of campylobacteriosis, 16 cases of EHEC, 17 cases of salmonellosis and less than 1 case of listeriosis (in a susceptible sub-population) when milk is consumed from the farm bulk milk tank. o 5 cases of campylobacteriosis, 49 cases of EHEC, 55 cases of salmonellosis and up to 17 cases of listeriosis (in a susceptible sub-population) when milk is consumed after farm gate sale. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 1 Raw milk can often be contaminated with pathogens, either directly through organisms shed as a result of udder infection or indirectly. Indirect contamination may arise from (i) a cow’s own faecal matter contaminating the udder and teats, (ii) faecal matter of other cows contaminating the udder (iii) milking clusters contacting surfaces with faecal contamination, and (iv) post-harvest environmental contamination. An intensive search of published and unpublished literature shows that internationally, raw cow milk is often contaminated with pathogens and, whilst data is scarce for Australia, the data which is available confirms that raw cow milk is a source of low levels of pathogenic microorganisms. Consumption of raw cow milk on school camps, during farm visits or via consumption of products marketed as pet or cosmetic milk, has been implicated in eight outbreaks of illness between 1998 2003 in Australia. Other than the burden of illness data, the quantitative modelling determined that: • Increased consumption of raw cow milk corresponds to an increase in the predicted number of illnesses • Inclusion of spoilage in the model resulted in an overall decrease in the number of predicted illnesses • Growth of pathogens occurred predominantly during domestic transportation and storage of raw milk • The occurrence and persistence of infection within herds or individual animals is often intermittent, which proves difficult to detect and routinely monitor • The ability to detect low levels of pathogens in raw milk is limited unless comprehensive sampling plans are used Understanding the extent to which cows in the herd carry pathogens in their gut, and shed them in their faeces, provides an important means of eliminating carrier animals and reducing the pathogen load in the farm and milking environment. Effective management of this would require high-level veterinary supervision and ongoing surveillance of individual animals in a herd. Such measures involve significant and often intensive interventions with concomitant enforcement authority oversight. Pathogen contamination of raw milk may be reduced by exercising enhanced hygienic control throughout the milk harvesting stage. Practices such as teat washing and dipping, foremilk stripping, and good milking hygiene will reduce the number of organisms (pathogenic and spoilage) that may enter the milk from environmental sources. For example, pre-milking udder washing with clean water and drying using hand towels reduces milk contamination by transient bacteria located on the exterior surfaces of the udder. Post-milking teat disinfection reduces the resident teat skin bacterial population, which is the main source of infection for the mammary gland. Test and hold practices, pending negative test results, is a commonly practised strategy for high risk, perishable foods. Raw milk is highly perishable, with both pathogenic and spoilage organisms expected to proliferate pending the receipt of test results. Furthermore, even low levels of pathogens in raw milk present a risk to consumers, and frequently these are below the levels of detection of current microbiological sampling regime prescribed in the Australia New Zealand Food Standards Code. Test and hold regimes may therefore be inadequate to determine with any confidence that contaminated raw milk will not reach consumers. Based on the available data, more than 200 sets of 5 x 25ml samples from every batch would be required to detect low level contamination in raw milk with 95% confidence. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 2 The safety of raw cow milk is influenced by a combination of management and control measures along the entire dairy supply chain. Control of animal health, adherence to good milking practices, and control over milking parlour hygiene are important in reducing the microbial load in raw milk. The modelling undertaken demonstrates that although the pathogen level may be very low in raw milk, there remains a risk of causing illness if consumed. The ability to detect pathogens in raw milk depends on the accuracy of testing, skill of personnel and the limit of detection for specific testing methodologies and pathogens. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 3 1. Background Food Standards Australia New Zealand (FSANZ) has responsibility for protecting the health and safety of consumers through the development of food standards. A comprehensive evaluation to identify and examine microbiological hazards along the entire dairy supply chain was conducted by FSANZ and is entitled A Risk Profile of Dairy Products in Australia1 (the Profile) (FSANZ, 2006). A key finding of the Profile was that Australian dairy products have an excellent reputation for food safety, as the majority of dairy products available for sale in Australia are made using pasteurised milk. This finding was supported by the paucity of evidence attributing foodborne illness to dairy products. The Profile did not specifically examine the risks to public health and safety associated with the consumption of raw cow milk; however it did confirm that internationally, unpasteurised dairy products are the most common cause of dairy associated foodborne illness. This document seeks to assess the risks to public health and safety resulting from consumption of raw cow milk. It utilises available scientific data and addresses uncertainty and variability in the conclusions drawn from the data. For example, the relevance and quality of data, the veracity of its source, and the assumptions made in the quantitative modelling are taken into consideration. The output of this risk assessment provides an estimate of risk to Australian consumers from specific pathogens following the consumption of raw cow milk. The outputs of the assessment will be used by FSANZ in the consideration of regulatory and/or non-regulatory measures as appropriate. 1 http://www.foodstandards.gov.au/_srcfiles/DAR_P296_Dairy_PPPS_Attach2%20Parts%20AB.pdf#search=%22Risk%20Profile%22 MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 4 2. Introduction 2.1 Purpose The purpose of this assessment is to provide an objective interpretation of available scientific data on the public health and safety risks associated with the consumption of raw cow milk and to provide advice on strategies which may be employed to impact on any identified risk. The assessment is undertaken in the context of the existing food safety management practices in the Australian dairy industry, and addresses the risk management questions: 1. What are the risks to public health and safety posed by the consumption of raw cow milk in Australia? 2. What are the factors that would have the greatest impact on public health and safety along the production chain for raw cow milk? The assessment considers specific microbiological hazards, and evaluates epidemiological and other scientific and technical data to determine whether these hazards have presented, or are likely to present, a public health risk in raw cow milk. The assessment also aims to identify where in the production and supply chain these hazards may be introduced, decreased or amplified. 2.2 Scope The scope of this risk assessment is to assess the risk to public health and safety from drinking raw cow milk. Assessing the risks resulting from consumption of further processed raw milk products, such as yoghurt and kefir, is outside the scope of this risk assessment. Although further processed products are outside the scope of this risk assessment, the findings may be used to determine the risks associated with the raw milk intended for processing into other raw milk products. The risk to public health and safety from consumption of selected raw milk cheeses has been undertaken separately2. 2.3 Definition of raw milk For the purposes of this assessment, raw milk is milk that has not been heat treated in accordance with the Australia New Zealand Food Standards Code3. 2 Microbiological Risk Assessment of Raw Milk Cheese http://www.foodstandards.gov.au/standardsdevelopment/proposals/proposalp1007primary3953.cfm 3 The Australia New Zealand Food Standards Code - Standard 4.2.4 – Primary Production and Processing Standard for Dairy Products MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 5 2.4 Approach Raw cow milk for human consumption is currently not permitted for sale in Australia; hence specific production and industry information is unavailable. It has therefore been assumed that at a minimum, practices, procedures and regulations pertaining to the existing bovine dairy industry would also apply to the production of raw cow milk. The assessment draws upon the findings of the Profile (FSANZ, 2006) which identified microbiological hazards in raw milk, primary production practices and other information relevant to raw cow milk. The assessment discusses microbiological hazards associated with raw cow milk, their attribution to raw milk mediated foodborne illness and the primary production and processing factors which impact on raw milk safety. The assessment is based upon the Codex risk assessment framework and utilises the outputs of quantitative modelling which estimate the risk per random daily serve of raw milk to consumers from enterohaemorrhagic Escherichia coli (EHEC), Listeria monocytogenes, Salmonella spp. and Campylobacter spp. present in raw cow milk. The model utilises data encompassing all stages from milking through to consumption. Hazard Identification Pathogens associated with raw cow milk Epidemiological data Hazard Characterisation Dose response Severity of illness Exposure Assessment Quantitative evaluation of exposure to Enterohaemorrhagic E. coli, Listeria monocytogenes, Salmonella spp. and Campylobacter spp. Prevalence and levels in raw milk at various stages: - Post-milking - Collection and packaging - Distribution and storage Consumption – frequency and volume Risk Characterisation Risk Estimates Figure 1: Main components of the microbiological risk assessment MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 6 3. 3.1 Findings Production of raw milk There is no formal marketing of raw cow milk for human consumption in Australia, although ongoing illegal sale of the product occurs in many jurisdictions. The system for producing raw milk would be expected to be similar to that used for producing cow milk for pasteurisation. However, the scale of operation would more likely be smallscale/boutique. Packaging of raw milk would be expected to occur on farm or at facilities close to the production environment with a capacity for segregating milk intended for raw or pasteurised product. The assessment identifies and describes the process and risk factors associated with production of cow milk. The main steps in raw milk production are described in Figure 2. STEP MILK PRODUCTION MILK COLLECTION MILK CHILLING AND IMPORTANT RISK FACTORS Animal production strategies including health status, housing and herd size Supplementary feed e.g. silage and water source Waste management Milking practices including teat washing and drying, stripping foremilk, etc Mastitis control measures Equipment cleaning and maintenance Rate and efficiency of chilling practices Equipment and personnel hygiene and sanitation STORAGE MILK PACKAGING TRANSPORT CONSUMER PRACTICES Equipment and personnel hygiene and sanitation Maintenance of chill temperatures Maintenance of chill temperatures Maintenance of chill temperatures during home storage Adherence to use-by-dates Time before consumption Figure 2: Generic steps in the production of raw cow milk MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 7 Measures employed to address the microbiological status of milk destined for pasteurisation would also apply to milk to be sold in the raw form. These include industry best practice regimes such as mastitis control programs, pre and post milking udder treatment practices and milking parlour hygiene programs. 3.2 Consumption of raw milk There is little data on the level of raw milk consumption in Australia or the potential demand. Generally, consumption data can be calculated from production statistics which provide information on the amount of a food commodity available to the population, or detailed information on the types and amounts of foods consumed by individuals via consumption surveys. For raw cow milk, this type of data is unavailable. It is probable that raw cow milk consumption is very low among the general population in Australia because it is currently not permitted and access is extremely limited. Even with limited availability, there are consumers of raw cow milk within Australia who purchase milk in the form of “pet milk” or via “cow share” programs. Unpublished research indicates that these consumers have very strong beliefs regarding the health benefits attributed to raw milk and their right to consume such products, and subsequently raw milk is their milk of choice. During 1994 in a survey of 3,999 respondents in the US, where raw milk sales are legal, only 3.2% reported consuming raw milk in the previous year (Headrick et al., 1997). Other studies undertaken internationally indicate generally low consumption of raw milk within the population. Consumption of raw milk is also practiced by dairy producers who have ready access to raw milk. While Australian data is limited, a study in the US found 42% of dairy producers (105/248 farmers in Pennsylvania) consumed raw milk and cited taste and convenience as their priority reasons (Jayarao et al., 2006). This is comparable to the results of Rohrbach et al. (1992) who found 34.9% of dairy producers consumed raw bulk milk. In the absence of reliable data for raw milk consumption within Australia, data on the consumption of pasteurised milk was used to estimate consumption of raw milk. It was not extrapolated to provide a population estimate. 3.3 Consumption of pasteurised milk Pasteurised milk and milk products are a significant component of the Australian diet. In 2006-07 Australians consumed around 2161 megalitres of market milk per annum (~ 100 litres per person per year), and this is expected to either remain static or decrease slightly (Dairy Australia, 2006). Milk consumption has been decreasing due to perceptions that it contributes to excess fat in the diet and the increased availability of substitute non-dairy products such as soybean-based drinks. Patterns of milk consumption have also been steadily changing from regular whole milk to modified milk types, such as reduced and low-fat milks and fortified specialty milks (Dairy Australia, 2004). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 8 Data obtained from the Australian National Nutrition Survey (NNS) in 1995 indicates that milk and liquid milk products comprise a significant component of the Australian daily diet (see Annex 1). Of 13,858 people surveyed, 11,311 (81.6%) reported consuming milk with an average intake across all consumers of 325g milk per day. The leading milk consumers were males aged 16 - 18 years, with a mean consumption of 495g/day, with a 95th percentile of 1,553g (even though they comprised only 1.5% of the surveyed population). Mean consumption of milk per day for children aged 2 - 3 years was 447g and 403g with 95th percentiles of 1,037g and 931g for males and females, respectively. Since 1995, a number of other surveys have been conducted to evaluate food consumption. The Roy Morgan Single Source Survey (RMSSS) is undertaken by an Australian market research company that monitors and provides information about the habits of over 50,000 Australian consumers. The data covers a vast range of consumers and indicates that more than 70 percent of the Australian population over 14 years of age consumes milk each 7 days, typically consuming over six glasses during this period. More detailed consumption information from both the NNS and RMSSS is contained at Annex 1. 3.4 Microbiological hazards in raw cow milk A wide range of pathogenic microorganisms have been found in raw milk, and many have been responsible for causing outbreaks of foodborne illness in Australia and internationally. In 1976 in Whyalla, South Australia, a large outbreak of Salmonella Typhimurium PT9 occurred involving over 500 cases. The majority (90%) of patients were under 15 years of age and most were under ten years. Characteristics of the outbreak suggested a common source and 95 percent of patients gave a history of consuming raw milk. Subsequent investigations isolated S. Typhimurium PT9 from 78 of the 273 persons investigated and 10 samples of bulk and bottle unpasteurised milk collected early in the outbreak (Seglenloka and Dixon, 1977). Raw milk was permitted for sale in South Australia at the time of the outbreak in 1976. Permissions for the sale of unpasteurised milk in South Australia were rescinded in 2000. An outbreak of the less common zoonotic pathogen, Streptococcus zooepidermicus was reported in Australia in 1992 involving three cases who had reportedly consumed unpasteurised milk from a house cow (Francis et al., 1993). Between 1998 and 2003, OzFoodNet’s Outbreak Register identified eight outbreaks comprising 101 cases of illness (and 4 hospitalisations) associated with the consumption of raw cow milk in Australia. Campylobacter spp. were the most common aetiological agent (5/8), with Cryptosporidium spp. and S. Typhimurium PT44 accounting for one outbreak each. There was one outbreak with unknown aetiology. Four outbreaks occurred on school camps where unpasteurised milk was consumed, while two outbreaks implicated unpasteurised milk consumed on farms. Unpasteurised milk was also consumed and led to outbreaks in a community setting and in a school. The outbreaks identified in the OzFoodNet Outbreak Register were investigated using three point source MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 9 studies and one case control study. Data prior to 2001 does not identify how outbreaks were investigated. Details of outbreak data from OzFoodNet’s Outbreak Register are listed in Annex 2. Internationally, consumption of raw cow milk has commonly been associated with foodborne illness. Between 1992 - 2000, 2% of all foodborne outbreaks in England and Wales were milkborne, with the majority (52%) attributed to the consumption of raw milk (Tayganyilmaz. et al., 2009). In the US where permissions for the sale of raw milk currently exist in 22 States, 58 outbreaks were attributed to raw cow milk between 1978 - 2000 (Tayganyilmaz. et al., 2009). More recently, outbreaks of S. Typhimurium occurred in the US in 2003 and 2007 involving 62 and 29 cases respectively. In California in 2006, E. coli O157:H7 was responsible for an outbreak involving 6 cases of illness with two cases contracting haemolytic ureamic syndrome (HUS). The median age of cases in this outbreak was 8 years and although geographically dispersed throughout California, all had consumed raw milk from the same dairy. E. coli O157:H7 was also responsible for an outbreak in 2005 in which 18 people became ill and four cases progressed to HUS. All patients were involved in a herd-share program and had consumed raw milk. Aetiological agents commonly associated with raw milk mediated illness include S. Typhimurium, E. coli O157:H7 and Campylobacter spp. Listeria spp. and other Salmonella serovars, including Dublin and Derby have also been reported. Outbreaks of illness associated with unpasteurised cow milk reported internationally are listed in Annex 2. A recent systematic review in New Zealand investigated the strength of evidence to support causal links between foodborne illness and consumption of raw milk products. The report findings indicated moderate evidence exists to support a causal link between consumption of raw milk products and Salmonella serovars, E. coli spp., Listeria monocytogenes and Campylobacter spp. (Jaros et al., 2008). In outbreak investigations sources of foodborne illness are generally determined through epidemiological and/or microbiological associations. The ability to identify an outbreak through the existing surveillance system is critical to enable an investigation to proceed. Difficulties exist in identifying and attributing illness to a particular food and include: • • • • • • • Food recall biases when gathering food consumption histories Time delays in recognition or notification of an outbreak Inability to trace food products to their source Reluctance of individuals to participate in investigations, particularly when they have purchased foods that are not permitted to be sold legally Long exposure windows for specific pathogens (e.g. Listeria monocytogenes) Inability to obtain representative food samples for analysis A lack of precision in or suitable methods for sample analysis and pathogen identification MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 10 It is important to recognise that outbreak data only represents a small proportion of actual cases of foodborne illness, as many outbreaks go unrecognised and/or unreported to health authorities. People do not always seek medical attention for mild forms of gastroenteritis, medical practitioners do not always collect specimens for analysis and not all foodborne illnesses require notification to health authorities. 3.5 Prevalence of pathogens in raw cow milk While a vast array of microbial pathogens may be encountered in raw milk, there is limited published data on the prevalence and levels of pathogens in raw cow milk in Australia. Milk processors screen incoming raw milk for a range of quality and shelf-life indicators but typically do not perform analyses for pathogens as the current practise of pasteurising milk destroys all pathogens. Where industry does collect such data it is rarely made public or published. A small survey conducted in Western Australia in 2007 (183 samples) found a high prevalence of organisms such as E. coli and coagulase-positive Staphylococcus aureus, whilst Salmonella spp. was reported at a prevalence of approximately 8%. No Listeria spp., Campylobacter spp. or EHEC were detected (Figure 3). Due to the limited number of samples and the restricted spatial and temporal conditions of the survey, insufficient data is available to draw conclusions on contamination levels within the entire Australian raw milk supply. The results do, however, indicate that raw milk produced under existing food safety management systems can contain pathogenic organisms which could result in illness if consumed. 100 90 80 70 60 Prevalence (%) 50 40 30 20 10 0 Campylobacter Coag Positive Staphylococci Figure 3: Coliforms E. coli EHEC Listeria Salmonella Prevalence of selected organisms in raw cow milk in WA (183 samples), WA unpublished (2007) raw milk data. A total of 28.4% of samples had Coliform concentrations of >100 cfu/ml, while 50.8% of samples had E. coli concentrations of >3 cfu/ml. The maximum concentration for both Coliforms (three samples) and E. coli (two samples) was greater than 25 000 cfu/ml. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 11 Published international literature indicates various pathogenic microorganisms may be associated with raw milk, including: Bacillus cereus, Brucella spp., Campylobacter spp., Coxiella burnetii, pathogenic E. coli, L. monocytogenes, Mycobacterium bovis, Mycobacterium avium subsp. paratuberculosis, Salmonella spp., S. aureus and Yersinia spp. (FSANZ, 2006). These include organisms shed by an infected animal (pathogens will be present in milk from mastitis animals) or organisms that enter the milk as a result of poor milking hygiene or from contaminated equipment and personnel. Table 1 provides a brief summary of the microbiological hazards, possible routes of contamination and the availability of epidemiological data. Table 1: Summary of microbiological hazards associated with raw cow milk Shed directly in milk# Severity of illness§ Implicated in foodborne illness × 9 Moderate ++ Severe^ ++ × 9 Severe^ + - + Severe^ + Enterohaemorrhagic E. coli × 9 Severe ++ Listeria monocytogenes 9 Severe^ ++ Salmonella spp. 9 Serious ++ Staphylococcus aureus 9 Moderate ++ Streptococcus spp. 9 - + Toxoplasma gondii 9 - ++ Yersinia enterocolitica 9 Serious + Organism Bacillus cereus Campylobacter jejuni/coli Clostridium perfringens Coxiella burnetii Cryptosporidium parvum Key: # ^ Transmission through udder; mastitis etc Susceptible sub-populations - § No data/unknown Based on ICMSF (2002) + Rare ++ More common Although C. burnetii infection has been associated with consumption of raw milk, ingestion is considered a minor route for human infection (Maurin and Raoult, 1999). Consequently little information exists regarding ingestion mediated illness. The causative link between Johnes Disease and Crohn’s Disease is tenuous. If there were a proven link, then the transmission of M. avium subs. paratuberculosis through the consumption of raw cow milk would be a risk. Tuberculosis resulting from milkborne transmission of Mycobacterium bovis has been drastically reduced in recent times worldwide by a combination of changing milk consumption habits, mandatory pasteurisation and cattle immunization programs (Ryser, 2001). Until recently, bovine brucellosis (Brucella abortus) was present throughout the world. A number of countries have succeeded in eradicating this disease including: Australia, Canada, Israel, Japan, Austria, Switzerland, Denmark, Finland, Norway, Sweden and New Zealand. B. melitensis remains endemic in southern Europe, west and central Asia, Mexico, South America and Africa. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 12 B. abortus in milk producing animals in Australia has been eradicated since 1989 and B. melitensis has never been detected in Australian herds. Australia has been recognised as bovine tuberculosis (M. bovis) free since 31 December 1997, and continues to conduct screening programs to monitor any M. bovis infection in dairy cattle. The Australian Quarantine and Inspection Service and Biosecurity Australia maintain import requirements focussed on animal health and biosecurity issues. Import conditions are currently being reviewed by Biosecurity Australia for Dairy Products, which includes consideration of Brucella spp. and Mycobacterium bovis. It should be highlighted that should these organisms be introduced into Australia through importation of contaminated raw milk products, they would pose a risk to consumers from consumption. While a range of pathogens associated with raw cow milk have been identified in the literature, this risk assessment only considers: Campylobacter spp., L. monocytogenes, E. coli (EHEC), and Salmonella spp. due to their likely occurrence in raw milk, their public health significance and access to suitable data to populate the quantitative model. M. bovis, M. avium subs. paratuberculosis, C. burnetii and Brucella spp. have not been further considered in this assessment. The frequency of contamination of raw cow milk with the specific pathogens evaluated in this assessment is described in Table 2. This data has been derived from a wide array of literature which is summarised in Annex 3. Table 2: Summary of prevalence data for pathogens in raw cow milk Organism Campylobacter jejuni Enterohaemorrhagic Escherichia coli (EHEC) Listeria monocytogenes Salmonella spp. International data 0 – 40% 0 – 33.5% 1 – 60% 0 – 11.8% Importantly, many of these pathogens may occur in milk concurrently. Jayarao et al. (2006) found 13% (32/248) of bulk milk samples contained more than one species of bacterial pathogen. Rohrbach (1992) reported a higher percentage (25%) of bulk milk samples contained one or more pathogenic bacteria. Inappropriate temperature control during the storage of raw milk following milking can lead to the growth of the majority of these pathogens. This may occur on farm, during transport, and packaging, and at various stages during marketing, including during transport, storage and in the home. 3.5.1 Campylobacter spp. Campylobacter spp. may be shed directly in the milk when the animal has clinical or subclinical mastitis due to Campylobacter infection, or indirectly through faecal contamination. Direct excretion of Campylobacter spp. into raw milk from mammary infection has been reported (Morgan et al., 1985; Orr et al., 1995). Campylobacter spp. have been isolated from faeces at prevalences up to 83% (Annex 3: Table 3.2). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 13 International data shows the prevalence of Campylobacter in raw milk varies between 1 – 40% (Table 2). From a small number of samples (36 samples) taken in South Australia during 1996 - 2000, no Campylobacter spp. were detected. Similarly, the Western Australian survey also detected no Campylobacter in 183 samples of raw cow milk during 2007. Campylobacter spp. are notoriously difficult organisms to culture because of the strict requirement for microaerophilic conditions, hence early studies (before the 1980s) frequently reported the absence of Campylobacter spp. in milk samples. More recently, the presence of non-culturable but viable Campylobacter spp. have been discussed (Sparks, 2009; Cools et al., 2005). 3.5.2 Enterohaemorrhagic E. coli Pathogenic E. coli are characterised into specific groups based on virulence properties, mechanisms of pathogenicity and clinical syndromes (Doyle et al., 1997). These groups include enteropathogenic E. coli, enterotoxigenic E. coli, enteroinvasive E. coli, enteroaggregative E. coli and enterohaemorrhagic E. coli (EHEC). Many synonyms are used to describe EHEC, including Shiga toxin-producing E. coli (STEC), Shiga-like toxin-producing E. coli (SLTEC), and verocytotoxin-producing E. coli (VTEC). These organisms are often found in the faeces of healthy cattle and as such their presence in raw milk is generally indicative of direct or indirect faecal contamination. However, organisms can be excreted through the udder when systemic infection results in mastitis. E. coli O157 (a particularly virulent strain of EHEC) has been isolated from raw cow milk both on farm and from bulk raw milk tankers (Desmarchelier and Fegan, 2003; Meng et al., 2001; Meng et al., 1998). E. coli O157 organisms have been found in many raw milk samples from around the world (ranging in prevalence from 1 - 33.5%) (Annex 3: Table 3.3). Contamination of E. coli O157 in Australian milk varies from 1 - 3% (Dairy Australia, 2006). The incidence of E. coli O157 in raw cow milk in Canada, USA, Europe and France has been reported as 2.3%, 3.2%, 3.6% and 2.4%, respectively (Schlesser et al., 2006). Faecal carriage of pathogenic E. coli has been reported at prevalences of up to 39% internationally. In Australia 68 faecal samples obtained in 2005 returned a prevalence of 12% (Annex 3: Table 3.4). E. coli O157:H7 produce verotoxins (VT) which affect Vero cells, hence they are known as verocytotoxigenic E. coli (VTEC) to reflect their biological activity. VTEC are also referred to as Shiga-toxin (ST) producing E. coli (STEC) or shiga-like toxin (SLT) producing E. coli (SLTEC). STEC strains produce two potent phage-encoded cytotoxins called Shiga toxins (Stx1 and Stx2) or verotoxins (VT1 and VT2). In addition to toxin production, another virulence-associated factor expressed by STEC is a protein called intimin. This protein is responsible for intimate attachment of STEC to the intestinal epithelial cells, causing attaching and effacing lesions in the intestinal mucosas. Intimin is encoded by the chromosomal gene eae, which is part of a pathogenicity island termed the locus for enterocyte effacement. Consequently, SLT, ST and VT have been used interchangeably, resulting in the terms verotoxin-producing E. coli (VTEC), shiga-like-toxin producing E. coli (SLTEC) and shiga-toxin-producing E. coli (STEC) coexisting in the scientific literature. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 14 In Australia, a range of EHEC isolates other than O157:H7 have been typically implicated in cases of HUS, including isolates of E. coli O157:H14, the non-motile E. coli O157:H-, E. coli O26:H11 and E. coli O111:H-. Testing for STEC is complicated due to the difficulty of recognising STEC in a background of commensal E. coli. STEC belonging to the O157 clone characteristically do not ferment sorbitol, distinguishing them from the vast majority of E. coli, hence colonies can easily be recognised on indicator media containing sorbitol. Culture on sorbitol MacConkey-based medium remains the most commonly employed method to screen for STEC. 3.5.3 Listeria monocytogenes L. monocytogenes is ubiquitous in the dairy environment; hence it may contaminate raw milk during milking. It can also be a cause of mastitis in milking animals and thus be shed directly into raw milk. Raw cow milk is often tested for L. monocytogenes and internationally, prevalence has been recorded up to 60% (Table 2). L. monocytogenes has not been reported as detected in either State monitoring programs, targeted surveys or in referenced literature in Australia. Carriage in the faeces of cattle has been recorded internationally at prevalence up to 50% (Annex 3: Table 3.6). 3.5.4 Salmonella Salmonella spp. can be found in the intestinal tract of most warm and cold blooded animals. In cattle, the bacterium are carried by both healthy and diseased animals, are shed in the faeces and hence can contaminate raw milk. Faecal carriage prevalence has been reported up to 36.4% in cattle (Annex 3 Table 3.8). International data shows the prevalence of Salmonella in raw cow milk ranging between 0 - 11.8% (Table 2). 3.6 Milk production and its impact on milk safety Primary production factors that impact on contamination and the microbiological status of raw cow milk have been comprehensively reviewed in the Profile (FSANZ, 2006). Pathogenic microorganisms may enter the milk directly via the udder of a diseased4 or infected5 animal or indirectly during the milking process due to contamination from the udder and teat canal, from the environment, from contaminated equipment, or from workers. The microbiological status of raw milk is influenced by animal health6, exposure to faecal contamination, environmental contamination and temperature control. The key risk factors that may affect the status of raw milk on-farm have been summarised in Table 3, along with strategies employed to manage the risk. 4 5 6 Disease is defined in the OIE Terrestrial Animal Health Code (2007) as the clinical and/or pathological manifestation of infection (http://www.oie.int/eng/normes/mcode/en_chapitre_1.1.1.htm) Infection is defined in the OIE Terristrial Animal Health Code (2007) as the presence of the pathogenic agent in the host (http://www.oie.int/eng/normes/mcode/en_chapitre_1.1.1.htm) Animal health is defined as incorporating both disease (the clinical and/or pathological manifestation of infection), infection and carrier status of the animal. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 15 Table 3: Key on-farm risk factors impacting on raw cow milk Risk factor Impact on milk safety Ways of managing the risk Disease Diseased milking animals will show increased shedding of pathogens directly into raw milk or faeces which may contaminate the production and milking environment. Infected animals with no signs of disease (asymptomatic carriers) may also harbour and shed pathogens, often intermittently, into milk and faeces. Animal health (including mastitis) control programs. Housing and husbandry Intensive housing practices may increase the risk of contamination of udders due to high stocking density, concentration of waste, stress and soiled bedding. Good herd management practices. Attention to animal welfare. Faeces Faeces may contaminate the exterior of the udder and introduce pathogens into raw milk. Reduce scouring. Udder hygiene at milking. Feed Contaminated or poorly prepared feed may increase faecal shedding of pathogens. Poor nutritional practices will affect scouring. Control over preparation, storage and distribution of feed, especially silage. Water Contaminated water used for stock drinking, teat washing and cleaning increases risk of environmental contamination. Ensuring water quality is suitable for purpose. Milking Poor milking practices, including dirty, chapped or cracked teats, inadequate cleaning and maintenance of milking equipment, and poor personnel hygiene can lead to contamination of raw milk. Pre and post milking udder emollients/antiseptics. Effective equipment maintenance, sanitation and cleaning practices. Storage Inappropriate temperature control of raw milk after milking can lead to growth of pathogens. Rapid cooling and holding of milk. Packaging/ Transport Packaging and poor hygiene may contribute to cross-contamination of raw milk. Inappropriate temperature control of milk during delivery can lead to proliferation of pathogens. Correct sanitising and packaging procedures Effective cold chain management. Defecation during milking may result in contamination of equipment or the generation of aerosols which may contaminate the milking environment (including milk storage vessels). Of the above risk factors, disease and milk have the major effect on milk safety. 3.6.2 Carrier Status Carriers are animals which are infected with the pathogen without exhibiting clinical signs of disease. These animals may have recovered from clinical disease or never had the disease. Their presence confounds conventional disease diagnosis and herd treatment and may result in the reappearance of a disease in a previously negatively tested group. Some carriers may be masked and not release organisms unless stressed or immunocompromised. The effects of stress and starvation on shedding of E. coli and Salmonella spp. in cattle have been known for decades (Grau et al., 1968; Grau et al., 1969). Detection of carrier animals is often difficult and may not be apparent until the infection re-activates. Destocking and complete replacement with disease free animals may be the only way of removing a disease carrier. The frequency and amount of pathogen excreted by a carrier varies with the organism, the animal, its husbandry and immune status, and the natural history of the disease in that animal species. In some diseases, carriers continue to be infected for many years while in others it can be a matter of a few months. Good husbandry will reduce stress but will not necessarily relieve certain types of production stresses such as pregnancy, parturition and lactation. These are significant stresses which modulate the immune system and can precipitate the excretion of organisms in a carrier animal. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 16 3.6.2 Pre and post-milking teat cleaning practices The teat surface is the major avenue of entry for microorganisms into raw milk. It is well recognised that significant opportunity exists for teats to become contaminated by faeces and soil (as dust or mud) (Cook and Sandeman, 2000; Vaerewijck et al., 2001). Pre-milking udder hygiene e.g. washing with clean water and drying using hand towels reduces milk contamination by transient bacteria located on the udder. Drawing foremilk ejects microorganisms which may have entered the teat canal. Post-milking teat disinfection reduces the resident teat skin bacterial population, which is the main source of infection for the mammary gland. The majority of Australian dairy farmers rely on post-milking teat disinfection, applied by spray techniques, as an integral part of their mastitis control programs (Lee, 1994). In dairy cattle, the rate of new intramammary infection due to S. aureus and Streptococcus agalactiae is reduced by approximately 50% when post-milking teat disinfection is practiced (Sheldrake and Hoare, 1980). 3.7 Post-milking factors impacting on milk safety On-farm cooling and hygiene practices are critical, with any failure adversely impacting the microbial load in raw milk. Correct sanitising procedures for packaging and effective cold chain management practices for the raw milk are important steps for minimising crosscontamination and growth of any microorganism present in the raw milk. Time and temperature conditions post-milking, i.e. through storage and distribution, have an important influence on the concentration of any contaminating pathogens. Even assuming the integrity of the cold chain is maintained, growth of psychrotrophic organisms such as L. monocytogenes and Yersinia enterocolitica can still occur at refrigeration temperatures if organisms are present in the milk. Other pathogenic microorganisms, if present, will also grow if the temperature increases by only a few degrees, i.e. E. coli and Salmonella spp. may grow at temperatures between 7 - 8oC (ICMSF, 1996). 3.8 Other factors impacting on milk safety A number of microbial inhibitors may be found in raw milk, including lactoferrin and the lactoperoxidase system, lysozyme, and specific immunoglobulins. Of these, the lactoperoxidase system (LP) system is the most effective. Lactoperoxidase is a naturally occurring enzyme in milk and in the presence of hydrogen peroxide and thiocyanate it has a bacteriostatic effect on many microorganisms and a bactericidal effect against some specific Gram-negative bacteria i.e. Pseudomonas spp. and E. coli. Natural levels of hydrogen peroxide and thiocyanate in raw milk are insufficient to activate the LP system and require addition. Once activated, its effect has limited duration which is influenced by the initial bacterial load, the species and strains of contaminating bacteria, and the storage temperature of the milk. An expert consultation on the LP system found it has a role to play as part of an integrated system to improve milk quality and safety. The system, however, is not considered a MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 17 replacement of existing well serving technologies, such as cooling and heat treatment. Instead it provides complimentary alternatives, particularly at the primary production stage when the other approaches are not available, feasible or suitable (FAO/WHO, 2006). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 18 4. 4.1 Quantitative Risk Assessment Introduction The risk assessment has identified four major pathogens in relation to raw cow milk which were considered in this quantitative model: Campylobacter spp., enterohaemorrhagic E. coli (EHEC), Salmonella spp. and Listeria monocytogenes. These four pathogens were chosen because of their likely occurrence in raw milk, their public health significance, and access to data to build a quantitative model. 4.2 Model overview The model is based on an unpublished model developed by the University of Tasmania and a paper by Clough et al. (2009). A modular approach has been taken by splitting the model into several discrete units: Farm level, Farm sales, Retail distribution and storage, Consumer distribution and storage and Consumption. Three scenarios were modelled: Scenario 1: A single serve (250 ml) exposure from the bulk milk tank Scenario 2: Domestic consumption after farm gate purchase Scenario 3: Domestic consumption after retail purchase7 Domestic consumption incorporates transport from the farm to the home and then storage in the domestic refrigerator. The overall structure of the model is illustrated in Figure 4 which depicts the various stages modelled for the production of raw cow milk. The model simulates farms with relatively small dairy herds that harvest and store raw milk under good temperature control. For Campylobacter spp., EHEC and Salmonella spp. the farm module considers a number of factors including with-in herd prevalence (i.e. the number of cows that carry any of the pathogens in their faeces), faecal pathogen concentration, degree of teat soiling, efficiency of teat cleaning prior to milking and individual cow milk yield. Information specific to dairy cows was used to develop the quantitative model inputs wherever possible. Instances where data from cattle at slaughter (e.g. faecal pathogen concentration) have been used as a surrogate have been indicated. The approach taken for L. monocytogenes is different to the other three pathogens as data on faecal concentration in dairy cows or beef cattle (as a surrogate) could not be found (Rhoades et al., 2009). To overcome this limitation, farm bulk milk tank prevalence and concentration data was used. 7 Retail purchase incorporates packaging, distribution and retail premise storage components. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 19 Farm Level Milk Production Contamination Microbial Growth Farm Sales Microbial Growth Retail Distribution and Storage Consumer Distribution and Storage Microbial Growth Consumption Figure 4: Flowchart of the raw milk model The potential for growth of any pathogens present in the raw milk is predicted at each stage between milking and consumption. The model also considers the growth of psychrotrophic Pseudomonas spp. to assess the effect of spoilage on the likelihood of illness. Secondary cases of illness due to person-to-person transmission are not considered in the estimation of risk. The quantitative model was built in Microsoft Excel® with simulations run using @Risk version 4.5.3 (Palisade Corporation). Twenty simulations each of 100,000 iterations were performed for each of the four pathogens, with the model outputs presented as predicted illnesses per 100,000 daily servings of raw milk. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 20 4.3 Model inputs Data used to develop inputs to the quantitative model were obtained from published and unpublished sources and expert opinion. A brief summary of how the data sources were used in the development of the quantitative model is presented below. 4.3.1 Bulk milk tank and with-in herd prevalence Scientific publications and unpublished sources reporting survey results for pathogen concentrations in bulk milk tank and with-in herd prevalence were evaluated against a number of exclusion criteria, including consideration of survey sample size, geographical location, animal age (weaners, heifers and cows) and microbiological methodology. A summary of the data sources for bulk milk tank and with-in herd prevalences that were accepted after review can be found in Annexes 5 and 6. A summary of the bulk milk tank prevalence for each of the four pathogens is presented in Figure 5. Each of the plots presents the prevalence of the pathogen ( = number of positive samples/total number of samples) with the error bars showing the 95th percentiles around the expected value (Newcombe, 1998). Statistical analysis of the bulk milk tank prevalence data was performed to investigate differences between published data and those from the most recent (unpublished) Australian survey data of raw cow milk from Western Australia (WA). This survey was limited in scope with five groups of samples taken during the autumn and winter months of 2007. A total of 183 raw milk samples were analysed for a range of indicator and pathogenic bacteria including generic E. coli, Campylobacter spp., EHEC, L. monocytogenes and Salmonella spp. A pairwise statistical test was performed between the WA bulk milk tank prevalence results and each of the results for the other studies. For Campylobacter spp. and EHEC, only one or two studies were found to have prevalences greater than the WA results. Results found to be statistically different to the WA results are highlighted with an asterisk (*) in Figure 5. The results for Salmonella spp. suggested that the WA results (14/183 detections) were greater than six other studies, predominantly from the UK. These six studies had much larger sample numbers (range 456 to 1720) than the WA survey. The UK Food Standards Authority surveys of raw milk in 2000 and 2003 had L. monocytogenes prevalences greater than the WA survey. No other results were found to be statistically different. There are several approaches that can be used to incorporate the bulk milk tank prevalence information of L. monocytogenes in the quantitative model. In each case a Beta distribution, Beta(s+1, n-s+1), where s is the number of positive samples and n is the total number of samples, is used. The simplest approach is agglomeration, where all of the positive results and the total number of samples from all surveys are summed. The resulting Beta distribution is Beta(Σs+1, Σn-Σs+1). This distribution does not include any of the structure of the individual studies. The second approach is to include each study in the model together with a weighting. This approach becomes unwieldy when many studies are considered. The third approach is an extension of the second where cumulative distributions for the Beta distributions for each of the studies are averaged (Vose, 2008). This approach captures the uncertainty within each study but the results, as a single cumulative distribution, is more readily included into the @Risk model. The mean bulk milk tank prevalence for L. monocytogenes using the cumulative distribution method is 4.1%. Although not used in this model, the prevalences for Campylobacter spp., EHEC and Salmonella spp. in bulk milk MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 21 tanks are 3.1%, 4.6%, 5.5%, respectively. These values do not include the data that were found to be statistically different to the WA survey results. The average cumulative Beta distribution approach was also used for the with-in herd prevalence results (see Annex 6). The mean with-in herd prevalence for Campylobacter spp., E. coli (EHEC), Salmonella spp. and L. monocytogenes was determined to be 25.6%, 6.4%, 8.9% and 5.8%, respectively. Ca mp yl o b ac ter spp. 0.20 0.20 EHEC 0.15 0.05 0.00 5 10 15 2 8 10 Listeria monocytogenes 12 0.20 ** * * 10 * * 15 0.00 0.05 0.10 Prevalence 0.15 0.20 S a l mo ne l l a spp. 0.15 * 5 0 5 Reference Figure 5: 6 Reference 0.10 0.00 4 Reference 0.05 Prevalence * 0.10 Prevalence 0.15 0.10 0.00 0.05 Prevalence * * 10 15 20 25 Reference Prevalence of Campylobacter spp., EHEC, Salmonella spp. and L. monocytogenes in bulk milk tanks. Error bars represent the 95% confidence intervals. Results highlighted with an asterisk (*) indicate that the prevalence is significantly different from the WA unpublished (2007) raw milk data. Details of each reference can be found in Annex 5. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 22 4.3.2 Pathogen concentrations in milk Pathogens can enter raw milk via a number of pathways including faecal contamination, foremilk, udder infection and environmental sources. Amongst these pathways, faecal contamination is often considered to be the most likely and the quantitative model has been developed for this route. Information necessary to model faecal transfer of pathogens into raw milk during milking will depend on a number of factors, including the number of lactating cows within a herd, milk yield for each cow, the with-in herd pathogen prevalence, the concentration of pathogens in faeces, the amount of faecal material present on the teat prior to cleaning, the effect of cleaning on reducing faecal material on the teat, and the amount of faecal material that is transferred. Each of these factors is considered below. Herd size, N For the quantitative model the herd size is assumed to be Log Normal with a mean of 30 lactating cows and a standard deviation of 10. A minimum herd size is set as 10 cows. Milk yield, Vi The daily milk yield is the amount of milk produced by a lactating cow. The yield will depend on a number of factors including the breed, time after calving, parity and health status (including mastitis). The total production of milk from a herd will also be influenced by calving practices. Farms that use seasonal calving will have a greater range in production between seasons than for those that calve throughout the year. An industry average of 20 litres, with a normal distribution and standard deviation of 5 litres, is used to characterise the annual average milk yield for an individual cow. With-in herd prevalence, pw The sources used to develop distributions for the with-in herd prevalence are presented in Annex 6. Concentration of pathogens in faeces, cf In a recent review of pathogens in cattle Rhoades et al. (2009) highlighted the limited number of published works reporting faecal pathogen concentration. As a result, survey data for cattle has been used as a surrogate for dairy cattle where necessary. It should be noted that cattle at slaughter will include a proportion of dairy cows. No data for concentration of L. monocytogenes in dairy cow faeces could be found. Faecal concentration data were obtained for generic E. coli and EHEC (Fegan et al., 2004c), Salmonella spp. (Fegan et al., 2004b) and Campylobacter spp. (Stanley et al., 1998). The logarithms of the faecal concentration were assumed to follow a Normal distribution. Where concentration data were missing a censored regression method was used. The mean (log10 cfu/g) and standard deviation for generic E. coli, Campylobacter spp., EHEC and Salmonella spp. are presented in the table below. Figure 6 presents a graphical summary of the concentration of generic E. coli, Campylobacter spp., EHEC and Salmonella spp. As might be expected the generic E. coli concentration is far greater than each of the pathogenic bacteria. The mean (base 10 logarithm) generic E. coli concentration is 5.9 compared with less than 1 for EHEC. The Campylobacter spp. faecal concentration is about 10 times greater than either EHEC or Salmonella spp. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 23 Bacteria Mean (log10 CFU/g) 5.90 1.79 0.87 0.75 0.4 Generic E. coli Campylobacter spp. EHEC Salmonella spp. Standard deviation (log10 CFU/g) 0.98 1.01 1.40 1.39 0.2 0.0 0.1 Density 0.3 C a m p yl o b ac te r spp. EHEC S a l m o ne l la spp. E .c ol i -4 -2 0 2 4 6 8 10 log10 Conc/g faeces Figure 6: Distributions for faecal concentration of generic E. coli, Campylobacter spp., EHEC and Salmonella spp. Amount of dirt present on the teat prior to cleaning, mf Vissers et al. (2006) performed experiments to estimate the amount of dirt ( = faecal, soil and bedding material) transferred from a teat during milking. The experiments used spores of butyric acid producing bacteria as a surrogate to estimate the amount of material transferred. Ten cows from each of 11 farms were used in the study. The results suggested that there were large differences between cows both in the amount of soiling of teats and the amount of material transferred from the teat surface into the milk. This variable is taken as the mass of faecal material transferred per litre of milk and not the total mass transferred as used in the model by Clough for EHEC in milk (Clough et al., 2009). Vissers et al. (2006) provided a table of the mean and standard errors for the concentration of spores on the udders of cows and the corresponding milk concentration. A preliminary experiment had established that there was no statistical difference in the spore concentration in the dirt on the udders or teats. This result is important as it is not possible to sample the teats prior to milking without influencing the transfer calculations. Using this evidence for the udder and milk concentrations plus the observation of a high correlation between these two concentrations (correlation coefficient = 0.79) it is possible to simulate the mass of dirt transferred per litre of milk. Analysis of this simulation revealed that the amount of dirt transferred was bi-modal; a result that may be interpreted as cows within a herd having either a low or high degree of teat soiling prior to teat cleaning and then milking. The mean amount of dirt transferred were 3.04 mg/l for cows with low teat soiling and 25.1 mg/l for cows with high teat soiling. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 24 Teat cleaning efficiency, ε Magnusson et al. (2006) studied the effect of a range of cleaning techniques for the removal of spores from teats. Alternative cleaning methods evaluated included the type of cleaning material (paper or cotton towels), chemicals, and cleaning duration. Cleaning efficiencies relative to the control ranged from 45 to 93%. Magnusson et al. (2006) suggested that the cleaning efficiencies for spores may be less than for bacteria due to their greater hydrophobicity. As such, the experimental cleaning efficiencies for spores may not be directly applicable to vegetative bacteria. A simulation approach is used to predict the cleaning efficiency by anchoring against Australian data on the concentration of generic E. coli in raw on-farm bulk milk tanks. Bulk milk tank concentration, cBMT (1) Campylobacter spp., EHEC and Salmonella spp. The approach taken to calculate the contribution of faecal contamination of teats is similar to that proposed by Clough et al. (2009). The Clough model was developed to investigate the effect of pasteurisation failure on the presence of EHEC in milk containers. The model was based on the information on teat cleaning for bacterial spores and not vegetative bacteria and used the total amount of faecal material transferred (Vissers et al., 2006) and not the mass per volume method outlined above. The final concentration in the bulk milk tank will depend on the factors outlined above. Each cow within the herd that have pathogens in their faeces may lead to teat contamination and subsequent transfer to the milk. The concentration of these pathogens in the milk will also be diluted by other cows in the herd that do not carry pathogens. In order to assess the possible teat cleaning efficiency for pathogenic bacteria it would be necessary to perform experiments similar to those performed by Vissers et al. (2007) but using vegetative bacteria. In the absence of such data an alternative simulation approach was investigated that incorporated the variables outlined above, together with the use of generic E. coli data from the WA raw milk survey. Generic E. coli was used as a surrogate as it is present at high concentrations in the faeces of dairy cows, is frequently isolated from raw milk, and quantitative data was available. The outputs of the simulation are presented in Figure 7. The horizontal axis is the logarithm of the generic E. coli concentration and the vertical axis is the cumulative proportion of herds tested. The simulation study considered a total of 30 herds. The solid data points in Figure 7 are the results for the WA survey for each of the five sampling periods. Individual sampling periods are not identified for clarity. The red 'staircase' to the right of the survey data is the simulation outputs assuming that there is no cleaning of teats prior to milking. In all but one case the simulation results are to the right of the survey data up to the 90th percentile. A trial-and-error approach was then used to determine a teat cleaning efficiency that shifted the simulated generic E. coli concentration to within the range of the survey data; the green 'staircase' on the left of Figure 7. It is apparent that the simulation does not fully capture the differences between survey periods, but it does shift the 'cleaned' teat results to the left by about one order of magnitude. This simulation study assumes that the only source of generic E. coli in raw milk is due to the faecal route. The teat cleaning efficiency was modelled as a Pert distribution with a minimum value of 0.9, a most likely value of 0.93 and a maximum value of 0.99. This teat cleaning efficiency is subsequently used for Campylobacter spp., EHEC and Salmonella spp. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 25 1.0 0.8 0.6 0.4 0.2 Cumulative proportion 0.0 Survey With teat cleaning No teat cleaning -1 0 1 2 3 4 −1 E c ol i conc (log10 CFU ml Figure 7: ) Comparison of the cumulative distributions for generic E. coli in raw milk: WA unpublished (2007) survey for raw milk (filled circles) and model predictions with (green lines) and without (red lines) teat cleaning MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 26 Mathematically the relationships between each of the variables in determining the bulk milk tank concentration are outlined below: N Vi VBMT herd size, N = Log Normal(30, 10) truncated at 10 volume of milk produced per day by the ith cow (litres) total daily herd milk production (litres) pw Nc ps cf mf with-in herd prevalence (-) number of dairy cows carrying faecal pathogens, Nc = Binomial(N, pw) probability that that the teats are lightly soiled, ps = 0.337 concentration of pathogens in faeces (cfu/g) teat cleaning efficiency, ε = Pert(0.9, 0.93, 0.99) mass of faecal material transferred from an unwashed teat per litre of milk (mg/l) T total number of pathogenic bacteria (as cfu) transferred from j positive cows cBMT Concentration of pathogenic bacteria in the bulk milk tank (cfu/ml) ε The method outlined does not explicitly include herd prevalence as a model input. Herd prevalence would be used to determine if a randomly selected herd has at least one cow with faecal carriage of pathogenic bacteria. A search of the literature could not identify suitable data sources for the three pathogenic bacteria, Campylobacter spp., Salmonella spp and EHEC. The unpublished University of Tasmania quantitative model used bulk milk tank prevalence as a surrogate for dairy herd prevalence. This approach, although appealing, has limitations due to the relatively high limits of detection for standard microbiological methods. As a result, the bulk milk tank prevalence will underestimate true herd prevalence. Not incorporating a herd prevalence for these three pathogens does have mathematical consequences in the running of the model. The combination of small herd sizes and low with-in herd prevalences would result in a substantial proportion of "positive" herds having no "positive" cows. The resulting risk estimates would therefore underestimate the risk to consumers. There are several approaches that can be taken to deal with this situation. The first is to ensure at least one "positive" cow is present in a "positive" herd. This would result in potentially higher effective with-in herd prevalence. An alternative is to not include the herd prevalence and to determine effective herd prevalence from those farms that do have at MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 27 least one "positive" cow with-in the herd. If sources can be found to establish the likely herd prevalence, then the risk estimates can be reviewed against the effective herd prevalence obtained from the quantitative model outputs. (2) L. monocytogenes Fenlon et al. (1995) described a survey of L. monocytogenes isolated from bulk milk tanks on Scottish farms over a 12 month period. The survey observed the occurrence and concentration of L. monocytogenes contamination of bulk milk tanks. Samples were taken at roughly monthly intervals and analysed using direct plating and enrichment techniques. Samples that were positive by direct plating had the count recorded, while samples positive after enrichment were recorded as presence only. The enrichment method has a limit of detection of 1 cfu/10ml. Censored regression was used to estimate the mean and standard deviation of the base 10 logarithm of the direct plated and enrichment results. The resulting distribution has a mean of 0.196 and a standard deviation of 0.677 and was used as the concentration for L. monocytogenes in the bulk milk tank. 4.3.3 Growth/inactivation rates The growth rate equations used to predict changes in the population of EHEC, Salmonella spp. and L. monocytogenes, plus the inactivation rates for the survival of Campylobacter spp. is provided in Annex 7. The growth rates for EHEC and Salmonella spp. were set to zero if the raw milk temperature dropped below 7 °C. Lag times were not explicitly included in the model formulation. 4.3.4 Dose response models Dose response models are used to link the number of pathogenic bacteria ingested (the dose) to the probability of an individual becoming ill. The dose response model for Campylobacter spp. was developed using results of a human clinical trial, while the models for EHEC, Salmonella spp. and L. monocytogenes were informed by epidemiological evidence. The end-point for the dose response models is illness in all cases. Beta-Poisson dose response models have been used for predicting illness for Campylobacter spp., EHEC, and Salmonella spp. There is no differentiation between the general population and susceptible sub-populations e.g. children for these pathogenic bacteria. The Campylobacter spp. dose response model is based on the human feeding trial reported by Black et al. (1988) and incorporates two stages. The first stage calculates the probability of infection after an exposure to a dose of Campylobacter spp., p(infection | dose), while the second stage determines the probability of illness occurring after the person becomes infected, p(ill | infected). The probability of illness after exposure is therefore: p(illness | dose) = p(infection | dose) x p(ill | infection) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 28 An α-β parameterisation of the Beta-Poisson dose response model is used: where log10 α = Normal(-0.767, 0.180) and log10 β = Normal(1.681, 0.742) with a correlation coefficient of 0.6455, and p(illness | infection) = Beta(8.855, 23.254). The dose response relationship for EHEC uses Shigella spp. feeding trial data as a surrogate. The use of Shigella spp. has been proposed in other risk assessments (see Marks et al. (1998)). The N50-α parameterisation of the Beta-Poisson dose response model was used for the Shigella spp. feeding trial data: 0.6 0.4 0.0 0.2 p(ill | Dose) 0.8 1.0 where log10 N50 is a discrete distribution with three equally weighted values of 2.38, 3.17, 4.48, and a fixed value of α = 0.266. A comparison of the Shigella spp. dose response models to the epidemiological evidence from EHEC illnesses from Strachan et al. (2005) indicated that the method was appropriate (Figure 8). 0 1 2 3 4 5 log10 Dose (CFU) Figure 8: Comparison of human EHEC outbreak data (Strachan et al., 2005) and the predicted dose response relationships based on human feeding trials for three strains of Shigella spp. The error bars represent the 95% confidence intervals for each outbreak. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 29 The Salmonella spp. dose response model is taken from Thomas et al. (2006). This model was based on epidemiological data for illnesses caused by a range of Salmonella serovars, predominantly S. Enteritidis and S. Typhimurium. An α−β parameterisation of the Beta-Poisson dose response model was used: where log10 α = Normal(-0.871, 0.89) and log10 β = Normal(1.727, 0.227) with a correlation coefficient of 0.892. For L. monocytogenes, two groups (susceptible and healthy) have been retained based on the epidemiological evidence highlighting the importance of susceptible populations and the occurrence of invasive listeriosis. Additional information on the importance of genetics of L. monocytogenes lineage has been explicitly included in the model as uncertainty. This uncertainty has been included in the values of the r-value used in the Exponential dose response model. The FAO/WHO (2004) estimated an average r-value for the susceptible population as 5.85x10-12. A more recent assessment of US epidemiological data which included genetic information regarding different L. monocytogenes strains found to cause invasive listeriosis, determined average r-values of 1.31x10-8 for lineage I and 5.01x10-11 for lineage II (Chen et al., 2006). These results suggest that there are large differences in virulence between L. monocytogenes strains. Severity has not been included for the risk estimates. 4.3.5 Consumption data In the absence of raw milk consumption data, data on pasteurised milk consumption from the National Nutrition Survey (NNS) was used. The NNS data was converted to total daily consumption volume as serving equivalents of 1 cup (250 ml) using a cumulative under-size method (i.e. 100 ml is treated as 250 ml) (Figure 9). The mean daily consumption for children and adults is 536 ml and 397 ml, respectively. The range of daily raw milk intakes was 250 - 1750ml. 70 Adults Children 60 50 Percentage (%) 40 30 20 10 0 <1 1-2 2-3 3-4 4-5 5-6 >6 Total Mammalian milk consumption (cups/day) Figure 9: Total mammalian milk consumption based on the NNS (1995) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 30 A summary of the model inputs is presented in Table 4. Table 4: Summary of quantitative model input distributions Input Distribution Bulk milk tank prevalence Campylobacter spp. Not applicable EHEC Not applicable Salmonella spp. Not applicable L. monocytogenes See Annex 5 Within-herd prevalence Campylobacter spp. See Annex 6 EHEC See Annex 6 Salmonella spp. See Annex 6 L. monocytogenes See Annex 6 Number of lactating cows in a herd Log Normal(30, 10) truncated at 10 Pathogen faecal concentration (log10 cfu/ml) Campylobacter spp. Normal(1.791, 1.007) EHEC Normal(0.865, 1.4) Salmonella spp. Normal(0.751, 1.39) Proportion of cows with lightly soiled teats 0.337 Concentration of faecal material transferred during milking from teat surface before cleaning (log10 mg/L) Lightly soiled Normal(0.483, 0.315) Heavily soiled Normal(1.399, 0.493) Teat cleaning effectiveness Pert(0.9, 0.93, 0.99) L. monocytogenes BMT concentration (log10 cfu/ml) Normal(0.1959, 0.6772) Maximum population density Spoilage bacteria (cfu/ml) 109 Pathogens (cfu/ml) 108 Spoilage criteria (cfu/ml) 10 7.5 On-Farm storage Bulk milk tank (BMT) time (h) Triangular(1, 16, 24) Bulk milk tank (BMT) temperature (°C) Triangular(2, 4, 10) Milk collection Milk tanker temperature (°C) Pert(4, 5, 6) Transportation time (h) Triangle(1, 3, 6) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 31 Table 4 cont: Summary of quantitative model input distributions Input Distribution Factory storage Factory storage time (h) Triangle(0, 48, 96) Factory storage temperature (°C) Pert(1, 4, 5) Production distribution Production distribution time (h) Triangle(0, 3, 24) Production distribution temperature (°C) Pert(0, 4.8, 7) Retail storage Retail storage time (h) Pert(12, 48, 96) Retail storage temperature (°C) Normal(4.9, 2.8) truncated at 0 and 6°C Domestic transportation Domestic transportation time (h) WeibullAlt(58%, 0.5, 93%, 1.5, "loc", 0) Domestic transportation temperature (°C) Pert(7.1, 16.25, 20.4) Domestic storage Domestic storage time (h) Cumulative distribution Domestic storage temperature (°C) Normal(3.88, 2.35) truncated at -2°C Raw milk consumption (ml/day) Children Empirical distribution Adults Empirical distribution Susceptible (for L. monocytogenes) Same as Adult Healthy (for L. monocytogenes) Same as Adult 4.4 Simulation results 4.4.1 Consumption scenarios The model outputs are reported as illness following consumption from the bulk milk tank (Scenario 1); domestic consumption after farm gate purchase (Scenario 2); and domestic consumption after packaging, distribution and retail sale (Scenario 3). A summary of the predicted illnesses in children per 100,000 daily servings of raw milk for Campylobacter spp., EHEC and Salmonella spp. is presented in Table 5. A summary of predicted illnesses in adults is presented in Table 6. The mean number of predicted illnesses for children in Scenario 1 (consumption of raw milk from the on-farm bulk milk tank) is 18.8 for Campylobacter spp., 16.5 for EHEC and 16.8 for Salmonella spp. per 100,000 daily servings. For Campylobacter spp. the mean number of predicted illnesses for children in Scenarios 2 and 3 decreases to 5 and <1 illnesses per 100,000 daily serves. This decrease in predicted cases is a result of the inactivation of Campylobacter spp. in chilled raw milk. The decrease in illnesses in Scenario 3 compared with Scenario 2 is due to the longer supply chain for raw milk that has passed through packaging, distribution and retail stages. The variability in predictions is in the order for 20%, as indicated by the 5th and 95th percentiles. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 32 For EHEC and Salmonella spp., the increase in the predicted number of illnesses for children between Scenario 1 and Scenarios 2 and 3 is due to the increased consumption and the growth that is possible between milking and consumption in the home. The predicted number of illnesses for children due to EHEC is 48.8 for raw milk purchased at farm gates sales and 96.9 per 100,000 daily servings for packaged raw milk sold at retail premises. The trend in the cases of illness per 100,000 daily servings for adults is consistent with those for children. However, the absolute values of illnesses for adults are lower than for children for Scenarios 2 and 3. This is as a result of the lower consumption amount recorded in the 1995 NNS. For example the mean number of illnesses due to Salmonella spp. for Scenario 3 is 130.2 for adults and 152.6 for children. As the dose response models for EHEC, Campylobacter spp. and Salmonella spp. do not include sub-populations, the number of illnesses for Scenario 1 is the same for both the children and adults. The small differences in values are due to random variations between simulations for each group. Table 5: Predicted cases of illness for children from Campylobacter spp., EHEC and Salmonella spp. per 100,000 daily serves of raw milk Pathogen Scenario Median Mean 5th percentile 95th percentile Campylobacter spp. 1 2 3 1 2 3 1 2 3 19 5 0 16 49.5 96.5 17.5 54 150.5 18.8 5.2 0.3 16.5 48.8 96.9 16.8 54.9 152.6 14 2 0 12 40 85 10 41 135 23 8 1 22 56 110 21 71 171 EHEC Salmonella spp. Table 6: Predicted cases of illness for adults from Campylobacter spp., EHEC and Salmonella spp. per 100,000 daily serves of raw milk Pathogen Scenario Median Mean 5th percentile 95th percentile Campylobacter spp. 1 2 3 1 2 3 1 2 3 19.5 4.5 0 16 38 77 15 61 128.5 19.9 4.7 0.1 16.5 38.4 77.7 14.9 59.4 130.2 15 2 0 12 34 66 8 46 117 24 8 1 22 46 89 21 69 148 EHEC Salmonella spp. The predicted illnesses for L. monocytogenes are presented in Table 7. For Scenario 1 the mean predicted number of illnesses is <1 for the three susceptible groups and the healthy sub-population. The highest result is for the susceptible sub-population when using the Chen Lineage I as indicated by the 95th percentile value of only 1 case of invasive listeriosis. The trends between Scenarios 2 and 3 are similar to the results for EHEC and Salmonella spp. A greater number of predicted illnesses are observed for Scenario 3 with a mean of 170 illnesses compared with 16.7 illnesses for Scenario 2 when using the Chen Lineage I r-value. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 33 Predicted cases of illness from L. monocytogenes per 100,000 daily serves of raw milk for the susceptible sub-population Table 7: Scenario 1 2 3 * Dose response parameter Chen Lineage I Chen Lineage II WHO Chen Lineage I Chen Lineage II WHO Chen Lineage I Chen Lineage II WHO Median Mean 5th percentile 95th percentile 0* 0* 0* 16 1 0* 166.5 5 1 0.2 0* 0* 16.7 0.9 0.1 170 5.6 0.7 0* 0* 0* 12 0* 0* 157.2 2 0* 1 0* 0* 23.7 2 0.85 190.55 10.85 2 Lower bound value (assuming <1 is equal to zero) There were no predicted illnesses when using the WHO dose response model for the general population. In contrast, using other dose response models, the number of cases of illness in susceptible sub-population scenario is between <1 and 170. The risk estimates in Tables 5 - 7 are made under the assumption that no spoilage occurs in the raw milk between milking and consumption. The inclusion of spoilage has the effect of eliminating the consumption of raw milk that has elevated levels of bacteria. The criterion for spoilage was exceeding a threshold of 107.5cfu/ml of a surrogate spoilage bacterium, Pseudomonas spp. If there is no temperature abuse and a long storage time then spoilage will occur with no, or minimal, corresponding increase in pathogen concentration. An alternative may be that spoilage occurs due to a period of temperature abuse that leads to the growth of a pathogen. The inclusion of a spoilage criterion may then remove the iterations where a large amount of pathogen growth has occurred. A summary of predicted illnesses for children incorporating spoilage is presented in Table 8. An example of the inclusion of spoilage is found for Salmonella spp. for Scenario 3. Results from the quantitative model determined that the mean number of illness per 100,000 daily servings of raw milk for children is 152.6 without spoilage and 65.9 with the inclusion of spoilage. Table 8: Predicted cases of illness for children from Campylobacter spp., EHEC and Salmonella spp. per 100,000 daily serves of raw milk, excluding spoilage Pathogen Scenario Median Mean 5th percentile 95th percentile Campylobacter spp. 1 2 3 1 2 3 1 2 3 19.5 4.1 0 16 31.4 46.6 15 48.6 63.9 19.9 4.7 0.1 16.5 32.7 47.1 14.9 47.3 65.9 15 3 0 12 25 40 8 36 54 24 7 1 22 40 59 21 54 81 EHEC Salmonella spp. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 34 4.5 Sensitivity analysis for on-farm factors Sensitivity analysis is a systematic evaluation of model inputs to identify the most important factors on the model outputs. A sensitivity analysis, using spider plots, was performed to assess those on-farm factors that have the greatest influence on the concentration of pathogens in the bulk milk tank. In the case of Campylobacter spp. four factors were considered: herd size, teat cleaning efficiency, teat soiling and with-in herd prevalence. Spider plots are constructed by replacing an input distribution by single values based on the cumulative percentiles (e.g. 1, 5, 25, 50, 75, 99) of the distribution. For herd size, the input distribution was a Log normal distribution with μ = 30 and σ = 10, truncated at 10. The corresponding values for the cumulative percentiles were 13, 17, 23, 28, 35, 49 and 61 cows. The 1st and 99th percentile with-in herd prevalence was 0.2% and 70.7%, respectively. The quantitative model was run for a total 10,000 x 7 x 4 iterations in order to construct the spider plot (Figure 10). Each axes of each panel in Figure 10 is the same; the x-axis is the percentile of the distribution and the y-axis is the mean Campylobacter spp. concentration calculated from the 10,000 iterations of the quantitative model. Input factors that have no influence on the model output can be readily identified as those that have horizontal spider plots. In the case of Campylobacter spp. concentration, the herd size had no influence, while the teat cleaning efficiency had only a weak influence (Figure 10). The two main factors that influence the mean Campylobacter spp. concentration are the degree of teat soiling and the with-in herd prevalence. 40 60 80 0 20 60 80 Percentile Teat soiling With-in herd prevalence 100 20 40 60 80 100 -5.5 -5.0 -4.5 -4.0 -3.5 Mean log10 CFU/ml 0 0 20 Percentile Figure 10: 40 Percentile -5.5 -5.0 -4.5 -4.0 -3.5 Mean log10 CFU/ml -4.5 -4.0 -3.5 -3.0 100 -3.0 20 -3.0 0 -5.5 -5.0 Mean log10 CFU/ml -4.5 -4.0 -3.5 -3.0 Teat cleaning efficiency -5.5 -5.0 Mean log10 CFU/ml Herd size 40 60 80 100 Percentile Spider plot for herd size, teat cleaning efficiency, teat soiling (solid squares: high soiling; open squares: low soiling) and with-in herd prevalence on the mean Campylobacter spp. concentration, as log10 cfu/ml MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 35 4.6 Measures to reduce levels of pathogens in raw milk The quantitative model has been developed to estimate the risk to consumers from the consumption of raw milk from a randomly selected farm, using three consumption scenarios. With no major interventions available to destroy pathogens in raw milk, it is essential to minimise the likelihood of pathogens contaminating the raw milk stream and then minimise the growth, if these risk estimates are to be reduced. Two ways in which to minimise contamination are to: • Minimise shedding of pathogens into milk and the milking environment • Minimise contamination during the milk harvesting stage Reducing pathogen shedding requires active surveillance of herds to reduce the numbers of carrier animals or those with sub-clinical infections. Herd prevalence data was used to determine if the milk from this ‘random’ farm was contaminated with any of the four pathogens. In all cases this prevalence was low, about 5%. Contamination levels in raw milk due to animals with intestinal carriage of pathogens are low, even when considering the variability about the mean concentration. For EHEC the mean concentration was estimated to be around 10-4cfu/ml (-4 log). As the limit of detection is 1cfu/ml, less than 2% of individual samples would be above this level and therefore EHEC would be rarely detectable using culture methods. When this probability is combined with the herd prevalence, the likelihood of successfully using random sampling of farm milk to detect animals shedding this pathogen is low. However, when considering the case of a single farm, additional factors need to be considered. One important factor is the persistence of pathogens within the animal and the farm environment. Persistence of L. monocytogenes infection on-farm has been examined by Fenlon et al. (1995). In this Scottish study, bulk milk tanks from 160 dairy farms were sampled at roughly monthly intervals over a 12 month period. All farms were tested four times and if a sample was positive then two additional samples were taken. The raw milk samples were analysed using direct plating (quantifiable) or by enrichment (detection only). Of the 160 farms tested, 25 farms were found to be positive at least once during the study period. Results for the 25 farms that were positive at least once, is presented in graphical form in Figure 4. The horizontal axis is the sample number and the vertical axis is the farm number. Note that all farms were tested at the start of the study (Sample 1) and again at sampling periods 4, 7 and 10. Samples positive by direct plating are shown as closed circles, those positive after enrichment as crossed circles and the no detects as open circles. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 36 Figure 11: Persistence of L. monocytogenes in farm bulk milk tanks The results of the study highlight large differences in the occurrence of L. monocytogenes between farms. For example, farms 1, 2 and 3 had detections in nearly all of the sample periods, while farm 25 only detected L. monocytogenes at the final quarterly sampling point and then it was not detected at the two subsequent sampling times. Similar behaviour can be seen with other farms. This suggests there may be important herd management factors that influence the introduction of L. monocytogenes into farms and its persistence in bulk milk samples. The second main means of reducing contamination is control over the milk harvesting stage, including pre and post milking teat cleaning regimes, foremilk stripping and milking hygiene. As discussed in Section 3.6.2, teat cleaning and disinfection are important for managing the introduction of microorganisms into the milk from the surface of the teat and the teat canal. This practice is also a significant control factor for managing mastitis in milking animals. Hygienic milking practices are firmly entrenched in current industry practices and are designed to reduce the risk of contamination and growth of microorganisms in the milk. Given that these practices are currently widely employed, there is little capacity for significant improvements to be made. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 37 4.7 Variability and uncertainty Variability and uncertainty in input variables have been included in all stages of the quantitative model development. This is especially the case for inputs that have the largest uncertainty. Other cases, such as uncertainty in the dose response model for L. monocytogenes, have also been included. Other inputs such as the concentration of Campylobacter spp., EHEC and Salmonella spp. include both variability and uncertainty in the development of the distributions used in the model. Model uncertainty and regression variability for the growth rate equations was not fully explored in this model, as these factors are generally of little importance as only limited growth of the pathogens was observed. The variability in the inactivation rate for Campylobacter spp. in raw milk was included in the quantitative model as this parameter was important in evaluating the risk from this pathogen. Australian specific data on pathogen prevalence for herds and for individual animals is scarce, as is any concentration data for pathogen levels within Australian raw milk. Consequently, international data has been used for these inputs. In the case of herd prevalence for EHEC, the majority of data has been obtained from surveys of herds in continental Europe and North America. The two Australian studies for the prevalence of EHEC in cows had mean prevalence’s of 9.5% and 0% compared with an average of 4.6% for the international studies. The inclusion of the international studies was to maximise the uncertainty for this important model input. The effective herd prevalence for Campylobacter spp., EHEC and Salmonella spp. were found to be 83%, 56% and 77%, respectively. As noted earlier, the herd prevalence was not included explicitly in the quantitative model for these pathogens. Clough et al. (2009) includes details of two studies that put the EHEC herd prevalence at 33.3% (95% confidence interval 21.7% - 45.0%) and 29.6% (95% confidence interval 15.1% - 49.8%). These results, taken with the observed variability between studies in the with-in herd prevalence data, suggest that for EHEC at least, the effective herd prevalence may not be too far from observed results. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 38 5. Discussion Prior to the introduction of mandatory pasteurisation in most Australian states in 1943, raw cow milk was implicated in widespread outbreaks of illness in Australia. Following its introduction there was a significant reduction in milkborne illnesses recorded; demonstrating pasteurisation was an effective means of eliminating pathogens from milk. Despite this, consumption of raw cow milk on school camps, during farm visits or via consumption of raw cow milk products has been implicated in eight outbreaks of illness affecting 101 people between 1998 – 2003. The majority of these incidents have involved milk consumed from a restricted distribution network, such as a single cow whilst on school camps or farm visits. The exception to this was in 1976, when over 500 people became ill after consuming raw milk which had been commercially produced and distributed. Campylobacter spp. and Cryptosporidium spp. were the most commonly implicated pathogens for the restricted distribution outbreaks, whilst Salmonella Typhimurium was involved in both a farm setting and the large retail outbreak. Internationally, consumption of raw milk continues to be a significant risk factor for foodborne illness, even when the milk is produced under regulatory control. Regular outbreaks of illness have occurred in the US in recent years. S. Typhimurium was implicated in outbreaks in 2003 and 2007, involving 62 and 29 cases of illness respectively. In 2005 and 2006, E. coli O157:H7 was responsible for outbreaks which lead to the development of six cases of haemolytic ureamic syndrome (HUS) With the 2006 outbreak, the median age was 8 years with the range between 6 and 18 years. There were six cases, two of which contracted HUS. The foregoing epidemiological evidence suggests the pathogens of most concern in raw cow milk mediated illness are Campylobacter spp., Salmonella spp. and EHEC. Listeria spp. are another commonly implicated pathogen in raw milk. While these pathogens are frequently implicated in foodborne illness, few studies have actually assessed the weight of evidence to support causal links between these pathogens and consumption of raw milk as a vehicle for foodborne illness. A systematic review undertaken by researchers in New Zealand concluded that moderate evidence existed to support causal links for raw milk mediated illness from Campylobacter spp. , Salmonella spp. , EHEC and Listeria spp. Consumption of raw cow milk appears to be low among the general population, but in specific groups large amounts are consumed, sometimes in the order of litres per person per day. Prevalent among this group of consumers is the belief that raw milk possesses particular healthy properties or attributes, above the existing nutritional components. As a result of these perceived health benefits, raw milk is often marketed to and consumed by individuals who may have lowered immunity such as the very young, very old or immunocompromised or to people with specific dietary needs. The very nature of how raw milk is produced and harvested means that raw milk can often be contaminated with pathogens. Contamination can occur either directly through organisms shed as a result of udder infections or indirectly. Indirect contamination may arise from (i) a cow’s own faecal matter contaminating the udder and teats, (ii) faecal matter of other cows contaminating the udder (iii) milking clusters contacting surfaces with faecal contamination, MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 39 and (iv) post-harvest environmental contamination. An intensive search of published and unpublished literature shows that internationally, raw cow milk is often contaminated with pathogens. Whilst data is scarce for Australia, the data which is available confirms that raw cow milk is a source of low levels of pathogenic microorganisms. In order to assess the risk to Australian consumers from consumption of raw cow milk, a quantitative microbiological risk assessment was undertaken. The assessment considered the risk associated with four identified key milkborne pathogens: Campylobacter spp., EHEC, Salmonella spp. and L. monocytogenes. The assessment determined predicted illnesses per 100,000 daily servings of raw milk consumed, at the bulk milk tank, in the home after farm-gate sale and in the home after packaging, distribution and retail sale. The first scenario of consumption of raw milk from the on-farm bulk milk tank is 19 cases for Campylobacter spp., 16 cases for EHEC and 17 cases of illness for Salmonella spp. per 100,000 servings for children. This scenario was developed to investigate a point source exposure e.g. a farm visit and represents a baseline risk estimate. In the case of Campylobacter spp. the likelihood of illness is lower when raw milk is consumed in the home after purchase from the farm-gate or from retail establishments. The quantitative model predictions show that the mean probability of illness for these two scenarios is 5 and <1 illness per 100,000 daily servings (Table 5). This reduction in illness is due to the finding that Campylobacter spp. does not survive well in refrigerated milk or other foods. The quantitative model is built on the assumption that EHEC, Salmonella spp. and L. monocytogenes are capable of growth in raw milk at rates determined by the storage temperature and time. The longer supply chain for raw milk sold at retail also results in a greater number of illnesses than for off-farm sales. A greater increase in predicted illnesses was seen for EHEC from Scenario 2 (49) to Scenario 3 (97) than from Scenario 1 to 2. Similarly, Salmonella spp. increased from 55 to 153 and Listeria spp. increased from 17 to 170 cases, respectively. The ability to reduce or minimise the risks associated with raw milk is considered to be quite limited. Enhancements in herd health and veterinary supervision, active surveillance to reduce the number of carrier animals, and improvements in milking hygiene could be expected to reduce the risk to consumers. However, they could involve significant and often intensive interventions, with concomitant intensive enforcement authority oversight. Another strategy would be to test and hold all raw milk before it is sold to consumers. Raw milk is highly perishable, and spoilage organisms would be expected to proliferate pending the receipt of test results. Furthermore, even low levels of pathogens in raw milk present a risk to consumers, and frequently these are below the detection levels for most microbiological screening methods. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 40 6. Data gaps and areas for further research During the preparation of the risk assessment a number of data gaps were identified along the raw cow milk supply chain. Research into these areas may better assist in describing factors along the farm-to-table continuum which impact on the burden of illness resulting from consumption of raw cow milk in Australia, enhance the modelling outputs, and reduce uncertainty. Significant data gaps exist hence further research is required in the following areas: • Spatial and temporal information on the prevalence and levels of pathogens in Australian dairy cows and in raw cow milk using the most sensitive detection methods • Data on the extent to which external faecal contamination on the udder and flanks etc can contaminate the milking environment and the milk • Carrier status of lactating cows for enteric pathogens and the effect this may have on the contamination level of raw milk • Information on the status of sub-clinical mastitis in lactating cows and the contribution this makes to the contamination levels in raw milk • Dose response models for pathogens • Pre and post milking teat treatments and their efficacy in removing contamination and improving the microbial status of raw milk • Raw milk consumption data, especially addressing the frequency and amount of raw cow milk consumed as well as the demographics of the consuming population MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 41 7. Conclusion Raw cow milk has always presented risks to public health because of the potential presence of pathogenic bacteria. The widespread adoption of pasteurisation for liquid milks was precipitated by ongoing outbreaks of illness, and the efficacy of the technique in eliminating most pathogenic microorganisms when combined with hygienic post-pasteurisation packaging and handling of the milk. This raw cow milk risk assessment sought to gather data on the public health impact of consuming raw cow milk, reviewed milk production and handling practices, and modelled the likelihood of illness associated with four key pathogens. While raw cow milk for human consumption is not legally available for sale in Australia, it is sold as “pet milk” or as “cosmetic milk” and is also made available to consumers during farm visits and farm stays. Between 1998 - 2003, OzFoodNet reported eight outbreaks comprising 101 cases of illness associated with the consumption of raw cow milk in Australia. Internationally, a large number of outbreaks of foodborne illness have been attributed to raw cow milk. The quantitative microbiological risk assessment undertaken during this assessment has determined that access to raw cow milk will result in an increased likelihood of foodborne illness in Australia. The burden varies depending on how the raw milk product is handled, with the risk for Campylobacter spp. greatest nearer the point where milk is procured from the cow, while the risks from EHEC, Salmonella spp. and L. monocytogenes increase as the milk is further handled, stored and transported. This reflects storage conditions and subsequent growth as the product passes along the supply chain. For raw cow milk the probability of illness after retail purchase is less than 1 case of campylobacteriosis, 97 cases of EHEC and 153 cases of salmonellosis in the general population for children, and as high as 170 cases of listeriosis in a susceptible sub-population. This is per 100,000 serves of a mean daily intake of 540 ml of milk to a child. The risk assessment also sought to identify ways of reducing the burden of illness. Measures available include addressing animal health, animal carrier status and milk procurement activities. Examples of additional controls may include regular testing of animals to determine if they are carriers of pathogens, exclusion of milk from animals with sub-clinical mastitis, advances in milking hygiene, and testing of milk before sale. The assessment determined that the capacity to make a significant impact on the burden of illness is limited. Current animal management structures are based on best practice for controlling mastitis and limiting the contamination of milk with spoilage and pathogenic microorganisms through the hygienic operation of the milking environment. Other measures such as testing of animals for carrier status, or analysis of raw milk are unlikely to detect low levels of pathogens. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 42 Annexes MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 43 ANNEX 1: Summary of milk consumption data Consumption data can be calculated from food production statistics or food consumption surveys. Food production statistics provide an estimate of the amount of specific food commodities available to the total population. This type of data may include national statistics on per-capita food production. Consumption surveys (such as national nutrition surveys) provide detailed information regarding the types and amounts of foods consumed by individuals or households and sometimes the frequency with which the foods are consumed. 1995 Australian National Nutrition Survey Data from the Australian National Nutrition Survey (NNS) provides detailed information regarding the types and amounts of dairy foods Australian’s are consuming. The most recent national survey was conducted during the period February 1995 - March 1996. Approximately 13,800 people aged two years or over from urban and rural areas in all States and Territories participated in the survey. Two approaches were used in the NNS to collect data on intake. The 24 hour recall method was used as the main indicator. All participants were interviewed by trained nutritionists who sought detailed information on all foods and beverages consumed during the day prior to the interview (from midnight until midnight). A sample of approximately 10% of the NNS participants also provided intake data for a second 24 hour period. A Food Frequency Questionnaire was used to assess usual frequency of intake for those aged 12 years or more. A summary of milk and liquid milk consumption by gender and age is outlined in Table 1.1. Table 1.1: Milk consumption by gender and age Gender Age (yrs) No. of Respondents No. of consumers (% of respondents) Mean milk consumed (g/day) 95th percentile (g/day) Male 2-3 170 150 (88.2) 447 1037 4-7 416 334 (80.3) 391 922 Female 8 - 11 385 313 (81.3) 395 903 12 - 15 349 269 (77.1) 452 1144 16 - 18 215 160 (74.4) 495 1553 19 - 24 485 356 (73.4) 341 1033 25 - 44 2140 1736 (81.1) 272 795 45 - 64 1554 1289 (82.9) 261 697 65+ 902 772 (85.6) 251 630 2-3 213 193 (90.6) 403 931 4-7 383 307 (80.2) 301 764 8 - 11 354 265 (74.9) 340 873 12 - 15 304 204 (67.1) 336 780 16 - 18 218 134 (61.5) 271 692 19 - 24 575 446 (77.6) 242 648 25 - 44 2385 2017 (84.6) 217 645 45 - 64 1752 1486 (84.8) 222 626 1058 880 (83.2) 217 557 13858 11311 ( 65+ MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 44 Roy Morgan Single Source Survey Roy Morgan data outlines the proportion of population groups who consume milk over a weekly period (weekly consumer) outlined in Table 1.2 below. Due to milk being a fairly staple commodity that can be consumed at various times throughout the day, it is likely that a 'weekly consumer' may indeed also be a 'daily consumer'. 'Milk' was defined in the survey to include low/no fat, regular and packaged/flavoured products. The number of respondents has been extrapolated to represent the total population (in 000’s) in that age group, based on Australian Bureau of Statistics data. Table 1.2: Roy Morgan Single Source Survey data Respondents Age (years) Gender 14 - 19 20 - 24 25 - 44 45 - 64 65+ Respondents Consuming Milk in past 7 days (%) 2005 2006 2005 2006 Male 879 899 660 (75%) Female 836 804 Male 671 Female Mean Glasses Consumed 2005 2006 660 (73%) 7.5 6.8 652 (78%) 618 (77%) 5.9 6.1 691 535 (80%) 490 (71%) 6.6 6.8 671 725 546 (81%) 603 (83%) 6.0 5.9 Male 2760 2766 2213 (80%) 2205 (80%) 6.7 6.8 Female 2852 2803 2321 (81%) 2274 (81%) 6.2 6.2 Male 2663 2690 2069 (78%) 2065 (77%) 6.1 6.3 Female 2938 3024 2329 (79%) 2356 (78%) 6.1 6.2 Male 1186 1233 857 (72%) 926 (75%) 6.1 6.0 Female 1112 1162 836 (75%) 853 (73%) 6.5 6.5 MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 45 ANNEX 2: Summary of foodborne illness associated with consumption of raw cow milk Australian data Outbreaks of foodborne illness attributed to the consumption of raw cow milk listed in OzFoodNet’s Outbreak Register are detailed in Table 2.1. Data has been compiled since 1995 with no outbreaks attributed to raw cow milk post 2003. Table 2.1: Outbreaks associated with milk and milk products in Australia, 1995 - 2004 (OzFoodNet data: 1995-June 2004) Aetiology Setting food prepared Comments 13 Campylobacter spp. School The risk of illness was 3.7 times higher among people who had consumed any unpasteurised milk. SA 14 Campylobacter spp. Camp Unpasteurised milk was supplied for drinks and cereal. 2001 Vic 12 Unknown Camp Relative risk was higher for those that had consumed unpasteurised milk. 2001 Qld 8 (4) Cryptosporidium spp. Community Strong epidemiological evidence of association between consumption of unpasteurised milk and cryptosporidium. 2000 SA 12 Campylobacter spp. Farm Infection due to consumption of unpasteurised milk. 2000 Vic 21 Campylobacter spp. Camp Infection due to consumption of unpasteurised milk. 1999 SA 12 Salmonella Typhimurium 44 Farm Infection due to consumption of unpasteurised milk. 1998 WA 9 Campylobacter spp. Camp Consumption of unpasteurised milk on camp. Year State 2003 Vic 2003 Cases (Hosp) Four outbreaks occurred on school camps where unpasteurised milk was consumed and two outbreaks were from unpasteurised milk consumed on farms. Unpasteurised milk was also consumed and led to outbreaks in a community setting and in a school. The outbreaks identified in the OzFoodNet Outbreak Register were investigated using three point source studies and one case control study. Data from before 2001 does not identify how outbreaks were investigated. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 46 International data Table 2.2: Outbreaks of illness associated with unpasteurised cow milk internationally Year Country Cases Causative Agent Cause (deaths) Reference 2007 USA 29 Salmonella Typhimirium Laboratory confirmed cased. Isolated from raw milk (MMWR, 2007b) 2006 USA 6 (2 HUS) E. coli O157:H7 Laboratory confirmed cased. Median age was 8 years (range 6 – 18). Geographically dispersed throughout California, all had consumed raw milk products from same dairy. (MMWR, 2008) 2005 USA 18 4 HUS E. coli O157:H7 Cow herd share program. 8/18 (MMWR, 2007a) laboratory confirmed. Isolated from raw milk, floor of milking parlour. Risk increased with increased consumption amount. 2003 USA 62 Salmonella Typhimurium Unpasteurised milk at dairy/petting zoo (Mazurek et al., 2004) 2003 USA 13 C. jejuni Unpasteurised milk (Peterson, 2003) 2001 USA 75 C. jejuni Unpasteurised milk obtained at a local (Harrington et al., 2001) dairy farm 2001 Austria 2 E. coli O157 isolated from dairy cow and goat, raw milk (Allerberger et al., 2001) 2000 Austria 38 C. jejuni Unpasteurised milk distributed by a local dairy (Lehner et al., 2000) 2000 Germany 31 C. jejuni Consuming raw milk farm visit (Thurm et al., 2000) 1998 Hungary 52 C. jejuni and C. coli Unpasteurised milk (Kalman et al., 2000) 1996 UK 33 C. jejuni resistotype 02 Educational farm visit, exposure to raw milk (Evans et al., 1996) 1995 USA 3 S. Typhimurium, Consumed in private home variate Copenhagen (CDC, 2002) 1993 USA 4 E. coli O157:H7 Consumed in a nursing home (CDC 2002) 1993 USA 6 E. coli O157:H7 Commercially distributed Unpasteurised milk (Keene et al., 1997) 1992 USA 50 C. jejuni Consumed at church (CDC 2002) 1992 USA 11 Campylobacter spp. Consumed in private home (CDC 2002) 1990 USA 13 Campylobacter spp. Consumed at school (CDC 2002) 1990 USA 5 Unknown Consumed in private home (CDC 2002) 1990 USA 42 C. jejuni Consumed at a dairy (CDC 2002) 1986 Austria 28 (5) L. monocytogenes Consumption of raw milk and biologically grown vegetables as possible source of infection (Allerberger and Guggenbichler, 1989) 1985 USA 25 C. jejuni Unpasteurised milk (Korlath et al., 1985) 1984 USA 23 not identified Associated with drinking raw milk from (MMWR, 1984a) local dairy 1984 Canada 9 C. jejuni A raw milk dairy (Anon., 1984) 1983 USA x 2 OB 31 26 C. jejuni A raw milk dairy (MMWR, 1983) 1983 USA 122 not identifed Associated with consumption of raw milk from a single dairy (Osterholm et al., 1986) 1983 USA ? (1) S. Typhimurium Unpasteurised milk (Tacket et al., 1985) 1983 UK 130 Streptobacillus moniliformis Unpasteurised milk (Shanson et al., 1983) 1982 USA 38 C. jejuni + thermotolerant strain (C. fetus subsp fetus) Unpasteurised milk (Klein et al., 1986) 19811983 USA 46 70 123(32) S. Dublin MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK (MMWR, 1984b) 47 Table 2.2 cont: Outbreaks of illness associated with unpasteurised cow milk internationally Year Country Cases Causative Agent Cause (deaths) 1981 USA 8 S. Derby Unpasteurised milk from a single dairy (Vogt et al. 1981) 1981 USA 250 C. jejuni Unpasteurised milk (Kornblatt et al., 1985) 1979 UK 700 S. Dublin Milk which had not been subjected to heat treatment (Small and Sharp, 1979) 1979 USA 3 C. jejuni All three patients had been consuming unpasteurised milk from a cow whose faeces were infected with C. fetus ss. jejuni (Blaser et al., 1979) 1973 1992 USA 40 outbreaks Various In states with legal raw milk (Headrick et al., 1998) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK Reference 48 ANNEX 3: Summary of key microbiological hazards associated with raw cow milk Organism Campylobacter jejuni/coli Shed directly in milk# 9 Severity of illness§ Implicated in foodborne illness Severe^ ++ ++ Enterohaemorrhagic E. coli 9 Severe Listeria monocytogenes 9 Severe^ ++ Salmonella spp. Key: 9 Serious ++ # ^ Transmission through udder; mastitis etc Susceptible sub-populations Table 3.1: Year Country 1982 USA - § No data/unknown Based on ICMSF (2002) + Rare ++ More common C. jejuni in raw milk Samples Prevalence (%) 195 1.5 0.9 Count Description - 1982 USA 108 1982 Netherlands 200 0 1982 USA 50 0 - UK 11 - F 5-R 40 - C 18.2 - F 40 - R 5-C - 1983 1984 UK 985 - R 153 - F 5.9 5.9 - 1987 USA 237 0.42 1987 UK 111 8.1 1988 Netherlands 904 4.5 192 - F 64 - D 1.56 - F 0-D - Reference 105 individual farm bulk tank milk. (Lovett et al., 1983) Bulk tanks of 9 Grade A farms. (Doyle and Roman, 1982) Farm milk tanks. (Oosterom et al., 1982) Milk from University dairy herd (Wyatt and Timm, 1982) F = bulk milk samples, R = pooled retail (Hutchinson et al., 1985) samples, C = individual cow samples. R = retail samples, F = samples from 12 (Humphrey and Hart, farm bulk milk tanks. 1988) Bulk tank trucks from 16 sites. (McManus and Lanier, 1987) 1 – 100 500ml farm bulk milk tanks – 111 MPN (Humphrey and Beckett, samples from five positive farms, 30 per 1987) samples from two negative farms. 100ml - Freshly drawn milk from individual cows, (Beumer et al., 1988) LP system inactivated. - F = 48 farms, D = 4 dairies. Samples collected over 4 x 1 month periods. Dairy samples from balance tanks, farm bulk (Davidson et al., 1989) milk tank coolers – enrichment procedure. - Herd milk samples tested by Federal Swiss Dairy Inspection and Advisory service. (Bachmann and Spahr, 1995) 1989 Canada 1990 Switzerland 496 0 1992 USA 292 12.3 - Individual farm bulk tanks – direct plate. (Rohrbach et al., 1992) 1993 Switzerland 3/83 3.6 - - (Wegmuller et al., 1993) 1.7 28 Retail outlets – samples ”green top” (De Louvois and milk. Rampling, 1998) 1996 UK 1097 1997 USA 131 9.2 1997 Canada 1720 0.5 1997 France 69 1.45 1999 UK 610 0.8 1999 Turkey 211 8.1 - Dairy plants. (Uraz and Yucel, 1999) 2001 Ireland 62 1.6 - On-farm bulk tank milk. (Whyte et al., 2004) Bulk tank milk. (Jayarao and Henning, 2001) - Samples from individual farm bulk tanks, (Steele et al., 1997) prior to tanker pick-up. - 27 different farms bulk tank milk over 6 (Desmasures, 1997) months, enrichment procedure - From 255 participating dairies, bulk tank (Food Standards Agency, milk. 2003) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 49 Table 3.1 cont: C. jejuni in raw milk Samples Prevalence (%) EU 1403 0.21 2001 US 248 2.2 2002 Pakistan 127 10.2 2002 EU 1431 0.35 2004 US 265 0 Year Country 2001 Table: 3.2: Year Country 1981 Count - Description Reference - (European Commission, 2003) Bulk tank milk from 248 dairy herds – (Jayarao et al., 2006) enrichment procedure - On farm bulk tank, enrichment, incubated (Hussain et al., 2007) 42 for 48h (92.4% C. jejuni, 7.6% C. coli) - (European Commission, 2004) Bulk tank milk samples (Murinda et al., 2004) C. jejuni in cattle Prevalence Description (% ) Samples Positive Reference USA 42 20 48 Rectal Faecal swabs from individual (Taylor et al., 1982) cows. 1982 Netherlands 200 11 5.5 Caecal samples slaughtered cattle. 1982 USA 50 0 0 Rectal swabs from University dairy (Wyatt and Timm, 1982) herd. 1982 USA 78 50 64 Rectal swabs. 1982 USA 15 2 13.3 1983 UK 40 18 45 In-milk samples. 1986 UK 12 10 83 12 herds, rectal swabs – prevalence of (Humphrey and Beckett, cows within herds ranged between 10 – 1987) 72%. 1988 Netherlands 904 197 21.8 13 farms, rectal swabs from individual (Beumer et al., 1988) cows. 2000 USA 2085 786 37.7 From 31 farms – individual cow direct (Wesley et al., 2000) faecal retrieval. 2002 USA 686 48 7 2002 USA 1450 745 51.4 Faecal cultures. 2002 USA 311 97 31.2 Dairy cattle – rectal or free faecal (Bae et al., 2005) samples. 2003 Australia 150 64 42.6 6 dairies - rectum retrieval. 2003 Finland 952 186 19.5 Rectal faecal samples, direct culture (Hakkinen et al., 2007) and enrichment procedure. 2004 USA 411 5 1.2 Faecal samples. 2004 USA 720 20 2.8 Lactating dairy cows, rectal faecal (Harvey et al., 2004) sample, 9 farms across the US. 2004 USA 1191 234 19.6 Cattle faecal samples from 60 farms. obtained Faecal samples taken implicated in outbreaks. from (Oosterom et al., 1982) (Doyle and Roman, 1982) from cows (Vogt et al., 1984) (Hutchinson et al., 1985) Faecal samples from dairy cattle sent to (Dodson and LeJeune, slaughter. 2005) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK (USDA/APHIS, 2003a) (Bailey et al., 2003) (Murinda et al., 2004) (Sato et al., 2004) 50 Table 3.3: E. coli in raw milk Year Country Samples Prevalence (%) Count per ml 1984 UK 985 – R 153 - F 63.4 – R 61.4 - F R = 11% >100cfu F = 2.6% >100cfu 1991 USA 23 4.3 - 1991 USA 115 10 1993 UK 329 0 77,172 4.4 1994 Trinidad 287 75.6 1994 Trinidad 507 4.9 - 1994 USA 603 0 1995 Trinidad 188 1997 France Isolate Description E. coli 3 year sampling period. (Humphrey and R = retail samples, F = bulk milk Hart, 1988) tanks from 12 farms. O157:H7 Raw milk samples collected (Wells et al., 1991) from 2 famrs during an outbreak investigation. O157:H7 Bulk raw milk samples from 69 different farms. - O157:H7 Study conducted over 15 (Mechie et al., 1997) months. 329 bulk milk tank samples. Samples were also taken from individual milk jars, milk filters and individual cows (fore and midstream). O157:H7 was isolated from one foremilk sample. - E.coli Milk samples submitted for routine testing. (Makovec and Ruegg, 2003) 13 (6.9%) of 188 strains of E. coli agglutinated with O157 antiserum (Adesiyun et al., 1995) VTEC Pre-processed pooled bulk raw milk from 16 milk collection centres. (Adesiyun, 1994) - O157:H7 Bulk milk tank samples obtained (Hancock et al., from routine testing laboratory. 1994) 6.9 - O157 Raw milk at 8 collection centres in Trinidad. (Adesiyun et al., 1995) 69 89.8 - E. coli 27 farms bulk milk tanks. (Desmasures, 1997) 1997 Scotland 500 0 - O157 Samples collected over 2 year (Coia et al., 2001) period from farm bulk milk tanks. 1997 USA 131 3.8 - STEC 131 farm bulk tank milk samples. 4 of 5 isolates of E. coli encoded for the Shiga toxin 2 gene, while 1 strain encoded for the Shiga toxin 1 gene. E. coli O157:H7 was not isolated. (Jayarao and Henning, 2001) 1997 Netherlands 1,011 0 - O157:H7 Bulk storage tanks originating from 1011 different dairy herds located throughout the Netherlands (Heuvelink et al., 1998) 1997 Canada 1,720 0.87 - VTEC Pick-ups (loads of raw milk from (Steele et al., 1997) a single farm bulk tank) from Ontario farm bulk tanks. Serotypes included O121, O26, O113, O163 1997 Trinidad 175 9.7 - VTEC Samples from collection centres. (Adesiyun et al., 1997) 1997 USA 42 0 - O157:H7 Samples obtained from 15 dairy (Ansay and Kaspar, processing plants. 1997) 1999 Northern Ireland 420 2.14 - VTEC Samples collected from 2 dairies. (McKee et al., 2003) 1999 Italy 100 0 - O157:H7 Bulk tanks of eight dairy farms. (Massa et al., 1999) 1999 UK 610 0.2 - O157:H7 From 255 participating dairies, bulk tank milk. (Food Standards Agency 2003) 1999 UK 610 52 From 255 participating dairies, bulk tank milk. (Food Standards Agency 2003) 1994 USA 4 4.2 x 10 - E.coli 6 1.6 x 10 cfu 2 18% = 10 - E. coli 5 10 cfu MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK Reference (Padhye and Doyle, 1991) 51 Table 3.3 cont: E. coli in raw milk Year Country Samples Prevalence (%) Count per ml Isolate Description Reference 2000 Italy 811 0 - O157 Samples obtained from dairy plant. (Conedera et al., 2004) 2000 USA 268 0.75 - O157:H7 O157:H7 - 30 dairy farms over one year period. (Murinda et al., 2002b) 6 33.3 10 cfu 3 ETEC Raw milk from three small-scale (Gran et al., 2003) dairies. ETEC positive samples produced heat-stable enterotoxin (ST1). 2001 EU and Norway 1629 3.4 - VTEC Summary zoonoses report. (European Commission, 2003) 2001 USA 248 2.4 - STEC Bulk tank milk from the 248 participating dairy herds. (Jayarao et al., 2006) 2002 Costa Rica 100 2 - O157:H7 Samples obtained from the (Reuben et al., principal producing zones of the 2003) country. 2002 EU and Norway 2968 0.7 - VTEC Summary zoonoses report. 2002 USA 859 0.6 - O157:H7 Bulk raw milk samples collected (Karns et al., 2007) during NAHMS Dairy 2002 study. Samples from 859 farms in 21 states analysed for EHEC virulence factors. - - 2000 Zimbabwe 2003 UK 0.35 +/E. coli 0.735logcfu - O157:H7 24 farms bulk tank milk. Farm bulk tank milk. Positive sample confirmed to be verotoxigenic, positive for VT2, eaeA and hlyA. (European Commission, 2004) (Hutchison et al., 2005) (De Rue et al., 2004) 2004 Belgium 143 0.7 2004 Malaysia 930 65 – E 33.5 - O 2004 Brazil 210 36.8 - E. coli Raw milk produced in 210 small (Nero et al., 2004) and medium farms located in four important milk-producing Brazilian states. 2004 Turkey 100 1 - O157 Raw milk samples collected (Oksuz et al., 2004) randomly from villages. Presumptive colonies confirmed serologically using antibodies to O157 antigen. 3 6.8 x 10 cfu E = E. coli Raw milk from 360 dairy farms sampled from 40 collection O= O157:H7 centres in 4 regions. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK (Chye et al., 2004) 52 Table: 3.4: Year Country E. coli in cattle Samples Positive Prevalence Description (% ) Reference 1993 USA 662 19 2.9 O157:H7 – Rectum retrieval of faecal samples (Zhao et al., 1995) from 50 control herds in 14 states. Counts 2 5 ranged from <10 cfu/g to > 10 cfu/g. 1993 Canada 291 49 16.8 VTEC – Rectal swab faecal samples from cows on 8 dairy farms testing positive in a previous study. (Rahn et al., 1997) 1993 UK 3593 153 4.3 O157:H7 – Study conducted over 15 months. Rectal swabs taken from lactating and nonlactating cows. (Mechie et al., 1997) 1994 USA 3570 10 0.28 O157:H7 – Faecal samples from dairy cattle in 5 of 60 herds. Samples were rectal swabs and (Hancock et al., 1994) fresh faecal pats. 1995 USA 965 31 3.2 O157:H7 - Study conducted as follow-up to the National Dairy Heifer Evaluation Project study. Direct rectum faecal retrieval of samples. (Garber et al., 1995) (Shere et al., 1998) (Garber et al., 1999) 1995 US 2395 105 4.4 O157:H7 - 14 month study on 4 dairy farms. 2 Counts ranged from 2.0 x 10 to 4 8.7 x 10 cfu/g. Samples obtained by digital rectal retrieval. 1996 USA 4361 52 1.2 O157 – Rectal retrieval of faecal samples collected from 91 dairy operations. 1996 Germany 726 131 18.0 STEC – 103 farms surveyed with 49.5% farms (Zschock et al., 2000) positive for STEC 1997 Australia 1802 171 9.5 STEC - Survey of Australian cattle. (Desmarchelier, 1997) 1997 USA 308 11 3.6 O157 – Samples from 19 dairy herds across three states obtained via faecal swabs. (Rice et al., 1997) 1997 Trinidad 313 68 21.7 VTEC – Faecal or rectal samples taken from dairy cattle. (Adesiyun et al., 1997) 1997 France 471 330 70 STEC – Faecal samples collected at slaughterhouse. Positive for stx gene. (Pradel et al., 2000) 1997 China 176 3 1.7 O157 – Faecal swab samples collected from two farms and two slaughterhouses. (Zhou et al., 2002) 1998 Australia 505 1 0.2 O157 – Faecal samples obtained from intestine post slaughter. (Hallaran and Sumner, 2001) 1999 India 206 37 18 STEC – Healthy domestic cow stool samples. (Khan et al., 2002) 1999 Australia 588 128 21.8 STEC – Samples collected on three occasions (Cobbold and over 12 month period. Rectal swabs. Desmarchelier, 2000) 2000 Korea 990 66 6.6 O157:H7 – Study conducted over 2 year period. Faecal samples obtained from beef and dairy cattle. (Jo et al., 2004) 2000 US 415 8 1.93 O157:H7 - 30 dairy farms over 1 year period. (Murinda et al., 2002b) 2000 Australia 128 12 O157:H 7 10 O26:H1 1 9.4 – O157:H7 7.8 – O26:H11 2000 Australia 310 39 12.6 O157:H7 - Faecal samples collected from cattle at slaughter. Three samples had counts 3 5 between 10 – 10 MPN/g. (Fegan et al., 2004c) 2000 Canada 240 2 0.8 O157:H7 – Fresh faecal samples from slaughterhouse pens, over a 1 year period (Van Donkersgoed et al., 2001) 2000 Denmark 1570 103 6.6 O157 – Eight dairy herds sampled four times over a 1 year period. Samples obtained by digital rectal retrieval. (Rugbjerg et al., 2003) O157:H7, O26:H11 – Faecal samples collected during a previous study and testing (Cobbold and positive for STEC were examined to determine Desmarchelier, 2001) virulence and serotype. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 53 Table: 3.4 cont: E. coli in cattle Year Country Prevalence Description (% ) Samples Positive Reference 2001 USA 305 16 5.2 O157 - Samples from 19 farms collected via rectal retrieval. Positive samples obtained from 7 out of 19 (36.8%) farms. (Cho et al., 2006) 2001 USA 720 52 7.2 O157:H7 - Faecal samples were obtained from lactating Holstein dairy cattle on four commercial farms in the south-western US. Samples obtained via rectal palpation. (Edrington et al., 2004) 2001 US 50 4 8 STEC – 4 farms sampled twice. Tested for stx (Cobbold et al., 2004) gene. 50 freshly passed faecal samples. 2002 USA 1026 21 2.1 O157:H7 - Isolated from cull dairy cattle at 2 (Dodson and LeJeune, livestock auctions in north-eastern Ohio. 1,026 2005) faecal samples were collected. 2002 USA 750 5 0.66 O157 - Comparison of mature dairy cattle in 50 dairy farms in Ohio and Norwegian using identical methodology. Samples collected via recto-anal swabs. (Lejeune et al., 2006) O157 - Comparison of mature dairy cattle in 50 dairy farms in Ohio and Norwegian using identical methodology. Samples collected via recto-anal swabs. (Lejeune et al., 2006) (Cho et al., 2006) 2002 Norway 680 0 0 2002 USA 132 6 4.5 O157 - Samples from 17 farms collected via rectal retreival. Positive samples obtained from 4 out of 17 (23.5%) farms. 2003 USA 574 118 20.6 O157:H7 – 144 cattle sampled four times by direct rectal retrieval over 7 month period. For (Khaitsa et al., 2006) each sampling period prevalence was 1.4, 6.9, 53.1 and 22.4%, respectively. 2005 USA 2043 527 25.8 STEC – 11 month period. 49 O serogroups were identified. (Renter et al., 2005) 2005 USA 408 16 3.9 O157:H7 - Samples collected over 24 month period from four different farms in four states. Rectal faecal swabs. (Doane et al., 2007) 2005 Brazil 100 39 39 STEC – Rectal swab faecal samples collected (Timm et al., 2007) at slaughterhouses. 2003 Australia 2005 Australia 22 108 - 13 - 12 O157 – Faeces of naturally infected cattle enumerated for O157 by a combination of MPN and AIMS. Counts of O157 between <3 4 MPN/g of faeces to 2.4 x 10 MPN/g. Comparison of E. coli and O157 counts. (Fegan et al., 2004a) O157 – 68 Faecal samples collected post evisceration and 40 samples collected from pen floors. Counts ranged from pen: <3 4 MPN/g to 4.6 x 10 MPN/g and from post 5 evisceration: 3.6 MPN/g to 7.5 x 10 MPN/g (Fegan et al., 2005a) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 54 Table 3.5: L. monocytogenes in raw milk Prevalence Count Description (%) per ml Year Country Samples 1972 Denmark 36199 1.2 USA 121 - M 14 - S 12 – M 14 - S - 1983 1985 USA 650 4.2 1985 Spain 95 1985 1992 Switzerland 1986 - Reference 23 year period testing bulk tank milk for herd prevalence in Danish mastitis control programme. 1132958 individual cows (quarter milk samples). (Jensen et al., 1996) M = Bulk tanks and milk collection trucks. S = milk filter socks at processing plant prior to pasteurising. (Hayes et al., 1986) - Farm bulk tanks in three areas of the USA. (Lovett et al., 1987) 45.3 - Over a period of 16 months. (Rodriguez et al., 1985) 4046 340 0.4 0.6 - Herd milk samples tested by Federal Swiss Dairy Inspection and Advisory service. (Bachmann and Spahr, 1995) USA 200 4 - Bulk storage tanks from 100 dairy farms at sampling times. (Liewen and Plautz, 1988) 1987 Finland 314 4.1 - Farm bulk tank milk from 80 farms. (Husu, 1990) 1987 USA 2511 2.9 - 19 dairy processing facilities tested bulk milk tankers. (Doores and Amelang, 1988) 1988 Canada 315 5.4 - Routine testing conducted over a 1 year period. (Slade et al., 1988) 1988 Canada 445 1.3 - Farm bulk tanks in three regions of Ontario. (Farber et al., 1988) - Bulk tank milk from 8 farms tested over 12 months. (Harvey and Gilmour, 1992) 1988 Ireland 113 5.3 1988 Australia 600 0 - Samples tested by NSW Dairy Corporation Laboratory. (Anon, 2003) 1989 Scotland 180 1.0 – 3.8 - 180 farm bulk milk tanks collected on 3 occasions. (Fenlon and Wilson, 1989) 1989 Germany 201 0 - Bulk milk tanks sampled during winter. (Eliskases-Lechner and Ginzinger, 1999) 1989 Canada 192 - F 64 - D 1.04 3.13 - 4 dairies and 48 farms collected over 4 x 1 (Davidson et al., month periods. Dairy samples from balance 1989) tanks, farm bulk milk tank coolers – enrichment procedure. 1989 France 2000 3.2 - Bulk tank milk used for producing raw milk cheese. (Sanaa et al., 1993) 1989 USA 300 3 - Bulk tank milk obtained from 12 farms over 13 month period. (Lund et al., 1991) 1989 Canada - - 1990 Canada 36 - T 36 – T 426 - B 2.8 - T 11.1 - T 1.9 - B - T = milk tankers, B = Bulk milk tank samples. (Fedio and Jackson, 1990) 1990 Australia 150 0 - Bulk milk tank samples from Queensland United Foods. (Ibrahim and Macrae, 1991) 1990 USA 292 4.1 Bulk tank milk sampled from 292 farms (Rohrbach et al., 1992) 1990 1991 Japan 943 - B 504 - F 0.32 - B 28.6 - F B = Farm bulk tanks from individual farms. Serotypes 1/2a and 4b.F = 1991 samples collected 9 times from 56 farms. (Yoshida et al., 1998b) 1991 Japan 51 50.9 3 farms bulk tank milk samples over 17 months. (Yoshida et al., 1998a) 1992 Canada 20 – B 401 - C 60 – B 5.2 - C 1992 England and Wales 2009 5.1 (Fedio et al., 1990) 4.6 x Identified during routine bulk milk testing, 1 3 10 cfu animal with subclinical mastitis in 1 quarter. No visible difference in milk. - - B = bulk tank milk, C = individual cow. Bulk tank (Fedio and Jackson, milk, individual cow milk and aseptic quarter 1992) milk from 4 farms. 62 cfu Farm bulk tanks sampled over 15 month period. (O'Donnell, 1995) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 55 Table 3.5 cont: L. monocytogenes in raw milk Year Country Samples Prevalence Count Description (%) per ml 1995 Scotland 640 6.6 35 cfu Bulk tank milk from 160 producers tested at 3 monthly intervals over 1 year. Prevalence over the intervals ranged from 4.4% – 9.4%. (Fenlon et al., 1995) 1997 Sweden 294 – B 295 - S 1.0 - B 19.6 - S 60 cfu B = Farm bulk tanks from 153 farms, sampled twice. (12 excluded from second sample set) S = Factory silos over 18 months. (Waak et al., 2002) 1997 USA 131 4.6 - 131 individual farm bulk milk tanks, all isolates contained O antigen (serotypes 1/2a, 1/2b and 1/2c). (Jayarao and Henning, 2001) 1997 Canada 1,720 2.7 - 3 day old samples from farm bulk tanks, collected prior to tanker pick-up. (Steele et al., 1997) 1997 France 69 5.8 - 27 individual farm bulk milk tanks. (Desmasures et al., 1997) 1997 France 1459 2.4 105 cfu 1998 Spain 774 3.62 - 25 farm bulk tanks over 50 month period. Reference (Meyer-Broseta et al., 2003) 114 farms collected bulk tank milk twice per season over 1 year period. (Gaya et al., 1998) 1998 USA 404 12.6 - In-line milk filters from 404 dairy farms. (Hassan et al., 2000) 1998 Turkey 100 4 - - (Vardar-Unlu et al., 1998) 1998 Mexico 1300 13 - 1300 samples of raw milk from 201 bulk milk tanks from 4 different dairies over a 1 year period. (Carlos et al., 2001) 1999 UK 610 17 1999 Brazil 12 8.3 1999 Turkey 2/211 0.94 2000 2001 2002 USA 474 474 25 4.9 7.0 68 - 3 States of Pacific Northwest. 474 herds samples over 3 time periods. 25 of 33 herds which tested positive in 2001 were tested in 2002. (Muraoka et al., 2003) 2000 Czech Republic 278 2.1 - 18 farms bulk tanks. (Navratilova et al., 2004) 2001 EU and Norway 1377 1.3 - Summary zoonoses report. (European Commission, 2003) 2001 USA 248 1.2 - Bulk tank milk from 248 dairy herds, enrichment (Jayarao et al., 2006) procedure. 2002 USA 861 6.5 2002 Slovak Republic 25 20 2002 Belgium 143 6.3 2002 US 860 6.5 2002 Brazil 210 0 2002 Portugal 105 1.9 From 255 participating dairies, bulk tank milk. ≤ 2log10 cfu - (Food Standards Agency 2003) 6 samples of raw milk from two different factory (Silva et al., 2003) receival areas. Serotype 4b. Dairy plant.s (Uraz and Yucel, 1999) Bulk tank milk samples from 21 different states collected during NAHMS Dairy 2002 study. 5 serotypes isolated - 1/2a, 1/2b, 3b, 4b and 4c. (van Kessel et al., 2004) - Raw milk samples. (Holko et al., 2002) - 143 farm bulk tank milk. Samples taken from cooled bulk milk. (De Rue et al., 2004) - NAHMS Dairy 2002 study – bulk tank milk of (USDA/APHIS, dairy operations with at least 30 milk cows over 2003b) 21 states. - 210 farms across Brazil using different milking systems. (Nero et al., 2004) - Samples obtained once a month from 7 processing plants over 1 year. (Kongo et al., 2006) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 56 Table 3.5 cont: L. monocytogenes in raw milk Year Country Samples Prevalence Count Description (%) per ml Reference 2003 Costa Rica 100 3.0 - Raw milk from principle producing zones of the country. (Reuben et al., 2003) 2004 Malaysia 930 1.9 - 360 dairy farms, milk collected at 40 collection centres. (Chye et al., 2004) 2006 Turkey 47 0 - Samples obtained from dairy plants. (Aygun and Pehlivanlar, 2006) Table: 3.6: L. monocytogenes in cattle Year Country Samples Positive Prevalence Description Reference 1992 Canada 69 14.5 21% Direct rectal retrieval of feacel samples (Fedio and Jackson, 1992) 1986 Finland 3878 373 9.6% Rectal faecal samples from 249 farms over two years (Husu, 1990) 1991 Japan 38 19 50% Three farms sampled over 17 months. Dropped faecal samples collected. (Yoshida et al., 1998a) Table 3.7: Year Country 1984 Salmonella spp. in raw milk Samples Prevalence (%) Count per ml Description Reference UK 2/985 0/153 0.2 0 - 985 retail samples, 153 samples from 12 farm bulk milk tanks. Serotypes were S. agama and S. Agona. (Humphrey and Hart, 1988) 1985 Canada 1140 2.5 - Collection of in-line milk filters. Serotypes included S. muenster, S. Newington and S. Montevideo. (McEwen et al., 1988) 1987 USA 678 4.7 - Bulk tank trucks from 16 sites. Serotypes included S. Montevideo (9), S. Bredeney, (5) S. Cerro (4) S. Infantis (3). (McManus and Lanier, 1987) 1989 Germany 201 0 - Bulk milk tanks sampled during winter. (EliskasesLechner and Ginzinger, 1999) 1992 England and Wales 1,673 0.36 - Farm bulk tanks sampled over 15 month period. Serotypes included S. Dublin, S. enteritidis, S. Typhimurium and S. Newport. (O'Donnell, 1995) 1992 USA 292 8.9 - Bulk tank milk sampled from 292 farms. (Rohrbach et al., 1992) 1993 Switzerland 456 0 - Herd milk samples tested by Federal Swiss Dairy Inspection and Advisory service. (Bachmann and Spahr, 1995) 1995 Canada 1,720 0.17 - Samples from individual farm bulk tanks, prior to tanker pick-up. Serovars isolated were S. muenster, S. thompson and S. Typhimurium. (Steele et al., 1997) 1997 USA 131 6.1 - 131 individual farm bulk milk tanks, Isolates of Salmonella belonged to group D (n = 4), B (n = 2), C (n = 1), and E (n = 1) “O” serogroups. (Jayarao and Henning, 2001) 1997 France 69 2.9 - Bulk tank milk from 27 different farms sampled over a six month period. S. Indiana isolated. (Desmasures, 1997) 1998 USA 404 1.5 - In-line milk filters from 404 dairy farms. 1 isolate was S. Typhimurium DT 104. (Hassan et al., 2000) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 57 Table 3.7 cont: Salmonella spp. in raw milk Year Country Samples Prevalence (%) Count per ml Description Reference 1999 UK 610 0.3 - From 255 participating dairies, bulk tank milk. One serotype was S. Goldcoast, the other unknown. (Food Standards Agency 2003) 2000 Ireland 29 3.4 - Compulsory monitoring programme. (Food Safety Authority of Ireland, 2004) 2001 USA 248 6.0 - Bulk tank milk from 248 dairy herds – enrichment procedure. Salmonella isolates were identified as S. Typhimurium (n = 10) and S. Newport (n = 5). (Jayarao et al., 2006) 2001 USA 268 2.2 - Bulk tank milk samples from 7 of 30 (25.3%) dairy farms were Salmonellapositive (Murinda et al., 2002a) 2002 USA 861 2.6 - Bulk tank milk samples from 21 different states collected during NAHMS Dairy 2002 study. Of the 22 positive samples, 9 serotypes were identified. Monetevideo (7), Newport (4) Muenster (2) Meleagridis (2) Cerro (2) Dublin (1) Anatum (1) and two others. (van Kessel et al., 2004) 2002 USA 854 11.8 - Samples collected as part of National Animal Health Monitoring System Dairy 2002 survey. (Karns et al., 2005) 2002 USA 860 2.7 - NAHMS Dairy 2002 study - bulk tank milk of dairy operations with at least 30 milk cows over 21 states. Most common serovars were Montevideo, Newport, Muenster, Meleagridis and Cerro. (USDA/APHIS, 2003b) 2002 Brazil 210 0 - 210 farms across Brazil using different milking systems (Nero et al., 2004) 2004 Belgium 143 0 - 143 farm bulk tank milk. Samples taken from cooled bulk milk. (De Rue et al., 2004) 2004 Malaysia 930 1.4 - 360 dairy farms, milk collected at 40 collection centres. Thirteen Salmonella serotypes were identified, including S. Muenchen, S. Anatum and S. Agona. (Chye et al., 2004) 1976 Australia - 10 Associated with an outbreak of S. Typhimurium phage type 9 in Whyalla. Salmonella isolated from bulk and bottled raw milk. (MMWR, 1977) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 58 Table: 3.8: Salmonella in cattle Year Country Samples Positive Prevalence (% ) 1991 USA 6862 145 2002 USA 415 2002 USA 2001 Description Reference 2.1 Conducted as part of the USDA National Dairy Heifer Evaluation Project, across 28 states. S. Typhimurium, S. Dublin most common serovars isolated. (USDA/APHIS, 1994) 9 2.17 Culled dairy cow samples. Faecal samples from 7 of 30 (25.3%) dairy farms were Salmonella-positive. (Murinda et al., 2002a) 585 39 6.7 Isolated from cull dairy cattle at 2 livestock auctions in north-eastern Ohio. 1,026 faecal samples were collected. (Dodson and LeJeune, 2005) Europe 25661 584 2.3 Summary zoonoses reports. (European Commission, 2003) 2000 Europe 27494 753 2.7 Summary zoonoses reports. Faecal samples obtained at slaughterhouse. (European Commission, 2004) 2002 USA 3669 267 7.3 USDA’s National Animal Health Monitoring System (NAHMS) conducted Dairy 2002, faecal samples were collected via rectal retrieval from approximately 5 operations in each of the 21 States participating in the study. (USDA/APHIS, 2003a) 2001 USA 720 262 36.4 Faecal samples were obtained from lactating Holstein dairy cattle on 4 commercial farms in the south-western US. Samples obtained via rectal palpation. (Edrington et al., 2004) 1991 USA 6861 145 2.1 Preweaned dairy heifers on 1063 operations tested. Faecal specimen obtained via rectal retrieval. (Losinger et al., 1995) 1996 US 91 25 27.5 4299 faecal samples from 91 herds. Faecal samples obtained via rectal retrieval. (Kabagambe al., 2000) 2002 Australia 310 21 6.8 Faecal samples from cattle presented for slaughter. Counts ranged from <3 MPN/gof 3 faeces to 2.8 x 10 MPN/g. (Fegan 2004b) et al., 2005 Australia 606 157 26 Counts obtained from cattle presented for slaughter. Maximum count in faeces was 93 MPN/g. (Fegan 2005b) et al., 1976 Australia 193 2 1 Detection associated with an outbreak of S. Typhimurium phage type 9 in Whyalla. (MMWR, 1977) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK et 59 ANNEX 4: 1. Hazard identification and characterisation of modelled pathogens Campylobacter spp. Campylobacter spp. are Gram-negative non-spore forming bacteria. Their cells are 0.2 - 0.8 μm wide and 0.5 - 5 μm long. They are mostly slender, spiral, curved rods, with a single polar flagellum at one or both ends of the cell. They are typically motile with a characteristic rapid darting corkscrew-like mobility (Smibert, 1984; Vandamme, 2000). Campylobacter spp. are classified under Campylobacteraceae, a bacterial family comprised of genera Campylobacter, Arcobacter and Sulfurospirillum. Among the 16 species and six subspecies of Campylobacter, two are most commonly isolated from stool samples of human gastroenteritis (Vandamme, 2000). They are Campylobacter jejuni subspecies jejuni and Campylobacter coli. C. jejuni accounts for approximately 95% of Campylobacter spp. caused human gastroenteritis, and C. coli are responsible for approximately 3 - 4% of the human illness. Campylobacter spp. are often a normal part of the intestinal flora of young cattle, sheep, goats, dogs, rabbits, monkeys, cats, chickens, turkeys, ducks, seagulls, pigeons, blackbirds, starlings and sparrows pigs (Smibert, 1984; Nielsen et al., 1997), and in blood and faecal material from humans with Campylobacter enteritis. They have also been found in the reproductive organs and oral cavity of humans and animals. Healthy puppies and kittens, rodents, beetles and houseflies have also been shown to carry Campylobacter spp. (Hartnett et al., 2002). Growth characteristics Campylobacter spp. require microaerophilic conditions for growth and have varying degrees of oxygen tolerance (3 - 5%) between species (Forsythe, 2000). Optimal growth occurs under conditions of 5% oxygen and 2 - 10% carbon dioxide (Park, 2002). Most strains do not grow in the presence of air, other than a few that may grow slightly under aerobic conditions. Some species can grow under anaerobic conditions with fumarate, formate and fumarate, or fumarate and hydrogen in the medium (Smibert, 1984; Vandamme, 2000). Campylobacter spp. grow optimally at 42 - 43oC. C. jejuni can grow in the temperature range of 30 - 45oC, pH of 4.9 - 9.5 and water activity above 0.99. At 32oC, C. jejuni may double its biomass in approximately 6 hours (Forsythe, 2000). Campylobacter spp. do not multiply at temperatures below 30oC, which means that the numbers of Campylobacter in foods will not increase at normal room temperatures (20 – 25oC). Although unable to grow below 30oC, Campylobacter remain metabolically active, are able to generate ATP, and are motile at temperatures as low as 4oC (Park 2002). Although Campylobacter spp. are considered thermotolerant, they are sensitive to heat and are readily inactivated by pasteurisation treatment or domestic cooking processes. Cooking at 55 - 60oC for several minutes readily destroys Campylobacter spp. The D value for C. jejuni at 50oC is 0.88 - 1.63 minutes (Forsythe, 2000). Campylobacter spp. are also sensitive to freezing and/or freeze thawing (Chan et al., 2001). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 60 Other than temperature, a range of other environmental factors including desiccation, oxidation and osmotic stress influences the survival of Campylobacter spp. They. are highly sensitive to desiccation and do not survive well on dry surfaces (Fernandez, 1985). The microaerophilic nature of Campylobacter spp. means that these organisms are inherently sensitive to oxygen and its reduction substances (Park 2002). Campylobacter spp. are much less tolerant to osmotic stress than a number of other foodborne pathogenic bacteria. For example, they are not capable of multiplication in an environment where sodium chloride concentration is 2% or higher (Doyle and Roman, 1982) Due to its sensitivity to environmental conditions and inability of growth at temperatures below 30oC or under aerobic conditions, the ability of Campylobacter spp. to multiply outside of an animal host is severely restricted. Although not capable of multiplication in food during processing or storage, Campylobacter spp. may have the ability to survive outside their optimal growth conditions (Park 2002). Pathology of illness C. jejuni causes fever and enteritis in human, resulting in acute inflammatory diarrhoea with clinical signs similar to those of other acute bacterial infections of the intestinal tract, such as salmonellosis. Principal symptoms are diarrhoea, nausea, abdominal pain, fever, myalgia, headache, vomiting and blood in faeces (Lastovica and Skirrow, 2000). The onset of symptoms is often abrupt with cramping abdominal pains quickly followed by diarrhoea. The mean incubation period is approximately 3 days with a range of 18 hours to 8 days. A particular feature of infection is abdominal pain, which may become continuous and sufficiently intense to mimic acute appendicitis. This is the most frequent reason for admission of Campylobacter enteritis patients to hospital (Skirrow and Blaser, 2000). Although incidents are rare, Campylobacter spp. have been implicated in causing a range of extra-intestinal infections including appendicitis, haemolytic ureamic syndrome, abortion, hepatitis, cholecystitis, pancreatitis, nephritis and others (Skirrow and Blaser, 2000). C. jejuni may cause septicaemia, meningitis and serious neurological disorders such as Guillain-Barré syndrome, an acute neuromuscular paralysis, and reactive arthritis such as Reiter syndrome (Lastovica and Skirrow, 2000). Mode of transmission Friedmann et al. (2000) examined data from 111 food and waterborne outbreaks of campylobacteriosis reported in the US between 1978 - 1996. Other than unknown foods, milk and water were the most common food vehicles associated with transmission of Campylobacter spp. Raw (unpasteurised) milk is largely responsible for dairy-related transmission. Of four milk-borne outbreaks in the period of 1990 - 1992, three were linked to raw cow’s milk and raw goat’s milk (CDC, 2003). Surveys in other developed countries, including the United Kingdom, Sweden, Germany, New Zealand, Denmark, US and Norway, indicate milk is the most frequent cause of foodborne Campylobacter spp. infection (Friedman et al., 2000). Outbreak data of foodborne campylobacteriosis recorded in Australia between 1992 - 2001 present a similar picture to the above, where approximately 42% of recorded outbreaks were the result of consumption of milk, and among this, raw milk accounted for approximately 80% of milk-borne Campylobacter spp. outbreaks. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 61 Published information by Eberhart-Phillips et al. (1997), Friedman et al. (2000), WHO (2000) and Vellinga and Loock, (2002) suggests that major routes of Campylobacter spp. transmission to humans are: • Consumption of food contaminated with Campylobacter spp., including consumption of raw and unpasteurised milk and milk products, consumption of undercooked meat such as poultry meat, and consumption of raw seafood • Consumption of water contaminated with Campylobacter spp. • Bathing or swimming in a Campylobacter spp. contaminated lake or pool • Direct contact with infected farm animals, such as cattle, sheep, chicken, etc • Contact with infected domestic animals, such as pet dogs, cattle and bird Incidence of illness C. jejuni is one of the most commonly reported aetiological agents of foodborne illness in developed countries, including Australia, NZ, UK and US (Mead et al., 1999; Park 2002). In the US, approximately 80% of all the cases of human campylobacteriosis are foodborne (Mead et al., 1999). In the period of 1998 – 2004, the notification rate of campylobacteriosis in Australia has been 100 – 120 cases per 100,000 population per annum. Notification rates were highest in the 0 – 4 year age group (Anon 2005). Occurrence in foods Foods potentially contaminated with Campylobacter spp. include raw and unpasteurised milk and milk products, raw poultry, raw beef, raw pork and raw shellfish, as well as foods that may have been exposed to water contaminated with Campylobacter spp. (Institute of Food Technologists, 2002). Virulence and infectivity of campylobacter Although not fully understood, Campylobacter spp. virulence is thought to involve production of microbial toxins. An enterotoxin Wassenaar (1997) abbreviated as CJT for C. jejuni toxin, is immunologically similar to the Vibrio cholerae toxin and the E. coli heat-liable toxin. At least six cytotoxins have been observed in Campylobacter spp., these being a 70-kDa cytotoxin, a Vero/HeLa cell cytotoxin, a cytolethal distending toxin (CDT), a shiga-like toxin, a haemolytic cytotoxin and a hepatotoxin. The CDT toxin has been shown to cause dramatic distension of human tumour epithelial cells, which leads to cell disintegration (Pickett et al., 1996). Active CDT toxin has been found in roughly 40% of the over 700 Campylobacter strains tested (Johnson and Lior, 1988). However, the role of enterotoxin and the cytotoxins in Campylobacter pathogenesis has not been fully identified. Dose response Dose-response relationships have been developed based on results from human feeding studies, whereby human volunteers were fed known numbers of Campylobacter spp. cells and then monitored for their response (Black et al., 1988). These models make the assumption that (1) a single cell has the ability to initiate an infection and (2) the probability of causing infection increases as the level of the pathogen increases. Data from human trial experiments indicates that Campylobacter spp. infection correlates proportionally to the dose ingested and MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 62 gradually reaches saturation. Despite a direct dose-response relationship being observed for the probability of infection, the probability of illness following from infection was independent of the dose ingested. The FAO/WHO Joint Expert Meetings on Microbiological Risk Assessment proposed a conditional probability of illness based on the probability of infection. Beta distribution of this conditional probability Hartnett et al. (2002) suggests the probability of illness is 20 - 50% after the establishment of a Campylobacter infection. For the human feeding trials 50% of individuals who ingested the minimum dose of 800 cells became infected (Black et al., 1988). Taking into consideration the limited size of the study, it has been proposed that the lowest infective dose would be somewhere close to 100 cells, which is comparable with epidemiological data (Prendergast et al., 2004) Immune status People with existing diseases are considered to have a higher susceptibility to campylobacteriosis than the general population (Pigrau et al., 1997). The incidence of Campylobacter spp. infection in patients with acquired immune deficiency syndrome has been calculated to be 40-fold higher than that in the general population (Sorvillo et al., 1991). In addition, 16% of Campylobacter spp. infections resulted in bacteraemia in these immunocompromised patients, a rate much higher than those occurring in the general population. References Anon. (2005) OzFoodNet: Foodborne disease in Australia: incidence, notifications and outbreaks. Annual report of the OzFoodNet network, 2004. CDI 29(2):164-190. Black, R.E., Levine, M.M., Clements, M.L., Hughes, T.P. and Blaser, M.J. (1988) Experimental Campylobacter jejuni infection in humans 1. J Infect Dis 157(3):472-479. CDC (2003) U.S. Foodborne Disease Outbreaks. http://www.cdc.gov/foodborneoutbreaks/us_outb.htm. Accessed on 16 March 2005. Chan, K.F., Tran, H.L., Kanenaka, R.Y. and Kathariou, S. (2001) Survival of clinical and poultry-derived isolates of Campylobacter jejuni at a low temperature (4 degrees C). Applied and Environmental Microbiology 67(9):4186-4191. Doyle, M.P. and Roman, D.J. (1982) Prevalence and survival of Campylobacter jejuni in unpasteurized milk. Appl Environ Microbiol 44(5):1154-1158. Eberhart-Phillips, J., Walker, N., Garrett, N., Bell, D., Sinclair, D., Rainger, W., B, a, t, e, s, , M and . (1997) Campylobacteriosis in New Zealand: results of a case-control study. J Epidemiol Comm Health 51(6):686-691. Fernandez, H. (1985) Desiccation resistance in thermotolerant Campylobacter species. Infection 13:197. Forsythe, S.T. (2000) The microbiology of safe food. Blackwell Science, Oxford. Friedman, C.R., Neimann, J., Webner, H.C. and Tauxe, R.V. (2000) Epidemiology of Campylobacter jejuni infections in teh United States and otehr industralized nations. In: Nachamkin, I. and Blaser, M.J. eds. Campylobacter. 2nd ed, American Society of Microbiology, Washington D.C., pp121-138. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 63 Hartnett, E., Paoli, G., Fazil, A., Lammerding, A., Anderson, A., Rosenquist, H., Christensen, B.B. and Nauta, M. (2002) Hazard identification, hazard characterization and exposure assessment of Campylobacter spp. in broiler chickens - Preliminary Report. Institute of Food Technologists (2002) Expert report on emerging microbiological food safety issues: implications for control in teh 21st enctry. http://www.ift.org/cms/?pid=1000379. Johnson, W.M. and Lior, H. (1988) A new heat-labile cytolethal distending toxin (CLDT) produced by Campylobacter spp. Microb.Pathog. 4(2):115-126. Lastovica, A.J. and Skirrow, M.B. (2000) Clinical significance of Campylobacter and related species other than Campylobacter jejuni and C. coli. In: Nachamkin, I. and Blaser, M.J. eds. Campylobacter. 2nd ed, American Society of Microbiology, Washington D.C., pp89-120. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M. and Tauxe, R.V. (1999) Food-related illness and death in the United States. Emerg.Infect Dis 5(5):607-625. Nielsen, E.M., Engberg, J. and Madsen, M. (1997) Distribution of serotypes of Campylobacter jejuni and C. coli from Danish patients, poultry, cattle and swine. FEMS Immunology and Medical Microbiology 19:47-56. Park, S.F. (2002) The physiology of Campylobacter species and its relevance to their role as foodborne pathogens. International Journal of Food Microbiology 74:177-188. Pickett, C.L., Pesci, E.C., Cottle, D.L., Russell, G., Erdem, A.N. and Zeytin, H. (1996) Prevalence of cytolethal distending toxin production in Campylobacter jejuni and relatedness of Campylobacter sp. cdtB gene. Infect.Immun. 64(6):2070-2078. Pigrau, C., Bartolome, R., Almirante, B., Planes, A.M., Gavalda, J. and Pahissa, A. (1997) Bacteremia due to Campylobacter species: clinical findings and antimicrobial susceptibility patterns. Clin.Infect.Dis. 25(6):14141420. Prendergast, M.M., Tribble, D.R., Baqar, S., Scott, D.A., Ferris, J.A., Walker, R.I. and Moran, A.P. (2004) In vivo phase variation and serologic response to lipooligosaccharide of Campylobacter jejuni in experimental human infection. Infect.Immun. 72(2):916-922. Skirrow, M.B. and Blaser, M.J. (2000) Clinical aspects of Campylobacter infection. In: Nachamkin, I. and Blaser, M.J. eds. Campylobacter. 2nd ed, American Society of Microbiology, Washington D.C., pp69-88. Smibert, R.M. (1984) Genus Campylobacter. In: Krieg, N.R. eds. Bergey's Manual of Systematic Bacteriology. Williams & Wilkins, Baltimore, pp111-117. Sorvillo, F.J., Lieb, L.E. and Waterman, S.H. (1991) Incidence of campylobacteriosis among patients with AIDS in Los Angeles County. J.Acquir.Immune.Defic.Syndr. 4(6):598-602. Vandamme, P. (2000) Taxonomy of the family Campylobacteraceae. In: Nachamkin, I. and Blaster, M.J. eds. Campylobacter. 2nd ed, ASM Press, Washington, DC, pp3-26. Vellinga, A. and Loock, F.V. (2002) The dioxin crisis as experiment to determine poultry-related Campylobacter enteritis. Emerg.Infect Dis 8:19-22. Wassenaar, T.M. (1997) Toxin production by Campylobacter spp. Clin.Microbiol.Rev. 10(3):466-476. WHO (2000) Campylobacter (Fact Sheet No 255). http://www.who.int/mediacentre/factsheets/fs255/en/. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 64 2. Escherichia coli (pathogenic) Escherichia coli are members of the family Enterobacteriaceae and are a common part of the normal intestinal flora of humans and other warm-blooded animals. The organisms are described as gram-negative, facultative anaerobic rod shaped bacteria (Desmarchelier and Fegan, 2003). Although most strains of E. coli are considered harmless, the species does contain certain strains that can cause severe illness in humans (Bell and Kyriakides, 1998). Strains of E. coli are differentiated serologically, based on O (somatic) and H (flagella) antigens (Lake et al., 2003). Pathogenic E. coli are characterised into specific groups based on virulence properties, mechanisms of pathogenicity and clinical syndromes (Doyle et al., 1997). These groups include enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), enteroinvasive E. coli (EIEC), enteroaggregative E. coli (EAEC) and enterohaemorrhagic E. coli (EHEC). Many synonyms are used to describe EHEC, including Shiga toxin-producing E. coli (STEC), Shiga-like toxin-producing E. coli, and verocytotoxin-producing E. coli. E. coli O157:H7 is the best known and most widely studied serotype of E. coli. One of its natural habitats is the intestines of cattle, which creates the potential for contamination of milk and dairy products. In spite of this risk, milk and dairy products have only occasionally been implicated in outbreaks of E. coli O157:H7 food poisoning, and even more rarely does an outbreak involve a pasteurised product (Kirk and Rowe, 1999). Growth characteristics Growth and survival of pathogenic E. coli is dependent on the simultaneous effect of a number of environmental factors such as temperature, pH and water activity. In general, pathogenic E. coli strains behave similarly to non-pathogenic strains, however certain EHEC strains have been found to have a higher tolerance to acidic conditions than other groups of E. coli (Desmarchelier and Fegan, 2003). The optimum temperature for growth of E. coli is 37°C, and it can grow within the range of 7 - 8°C to 46°C (ICMSF, 1996). Heat sensitivity of pathogenic E. coli is similar to that of other Gram-negative bacteria and is dependent on the pH, water activity and composition of the food (Bell and Kyriakides, 1998). Due largely to its importance as a cause of foodborne illness in the US, most studies on the growth and/or survival of pathogenic E. coli have been undertaken with E. coli O157:H7 (an EHEC organism). Studies on the thermal sensitivity of E. coli O157:H7 have revealed that it is no more heat sensitive than Salmonella (Doyle and Schoeni, 1984). Therefore, heating a product to kill typical strains of Salmonella will also kill E. coli O157:H7. Studies have demonstrated that some EHEC strains are acid-tolerant and can survive for at least five hours at pH 3.0 - 2.5 at 37°C (Benjamin and Datta, 1995). Stationary phase and starved pathogenic E. coli have been found to have an increased acid tolerance compared with exponential growth phase organisms (Arnold and Kaspar, 1995). Pathogenic E. coli may therefore be able to survive and/or grow in food products previously considered too acidic to support the survival of other foodborne pathogens. The effect of pH on E. coli survival is, however, dependent on the type of acid present. For example, E. coli O157:H7 can survive in a medium adjusted to pH 4.5 with hydrochloric acid but not when adjusted to the same pH with lactic acid (ICMSF, 1996). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 65 The minimum water activity required for growth of pathogenic E. coli is 0.95, or approximately 8% sodium chloride (ICMSF, 1996). In sub-optimal temperature or pH conditions, the water activity required for growth increases (Desmarchelier and Fegan, 2003). Pathology of illness EPEC causes illness primarily in infants and young children in developing countries. Symptoms include watery diarrhoea, with fever, vomiting and abdominal pain. The diarrhoea is usually self-limiting and of short duration, but can become chronic (more than 14 days). EPEC is also recognised as a foodborne and waterborne pathogen of adults, where it causes severe watery diarrhoea (with mucus, but no blood) along with nausea, vomiting, abdominal cramps, fever, headache and chills. Duration of illness is typically less than three days (Doyle and Padhye, 1989; Dalton et al., 2004) ETEC is another major cause of diarrhoea in infants and children in developing countries, as well as being recognised as the main cause of ‘travellers diarrhoea’ (Doyle and Padhye, 1989). Symptoms include watery diarrhoea, low-grade fever, abdominal cramps, malaise and nausea. In severe cases the illness resembles cholera, with severe ‘rice-water’ diarrhoea and associated dehydration. Duration of illness is 3 - 21 days (Doyle and Padhye, 1989). EIEC cause a dysenteric illness similar to shigellosis. Along with profuse diarrhoea, symptoms include chills, fever, headache, muscle pain and abdominal cramps. Onset of symptoms is usually rapid (<24 hours) and may last several weeks (Doyle and Padhye, 1989). EHEC infection normally results in diarrhoea-like symptoms. Haemorrhagic colitis, an acute illness caused by EHEC organisms, is characterised by severe abdominal pain and diarrhoea. This diarrhoea is initially watery but becomes grossly bloody. Symptoms such as vomiting and low-grade fever may be experienced. The illness is usually self-limiting and lasts for an average of 8 days. The duration of the excretion of EHEC is about one week or less in adults, but it can be longer in children (ICMSF, 1996). Complications resulting from EHEC infections vary. About 5% of haemorrhagic colitis victims may develop haemolytic ureamic syndrome (HUS) (European Commission, 2000). This involves the rupture of red blood cells (haemolysis), subsequent anaemia, low platelet count and kidney failure. The case-fatality rate of HUS has been reported to be 3 – 7% (Codex, 2002). Shiga toxins produced by EHEC attack the lining of the blood vessels throughout the body, predominantly affecting the kidney. However other organs such as the brain, pancreas, gut, liver and heart are also affected and may result in further complications such as thrombotic thrombocytopenic purpura. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 66 Table 1: Clinical, pathological and epidemiological characteristics of disease caused by the five principal pathotypes of E. coli (Robins-Brown, 1987) Pathotype Clinical symptoms Intestinal pathology Susceptible population ETEC Watery, cholera-like diarrhoea No notable change Children in developing countries; travellers to those countries EIEC Bacillary dysentery Inflammation and disruption of the mucosa, mostly of the large intestine All ages; more common in developing countries EPEC Non-specific gastroenteritis Attaching-effacing lesions throughout the intestine Children under 2 years of age in developing countries EHEC Bloody diarrhoea “Haemorrhagic colitis”; attaching-effacing lesions confined to the large intestine; necrosis in severe cases Children and the elderly in developed countries. EAEC Persistent diarrhoea Inflammation, cytotoxic changes in enterocytes (data from experimental studies) Children in developing countries; travellers to those countries Mode of transmission Pathogenic E. coli are transmitted by the faecal-oral route. Sources of transmission include person-to-person, foodborne, waterborne (drinking water and direct contact with faecal contaminated water) and direct contact with infected animals (ICMSF, 1996). Incidence and outbreak data Infection with pathogenic E. coli is a cause of significant morbidity and mortality worldwide. Outbreaks caused by EPEC, ETEC and EIEC occur infrequently in developed countries (ICMSF, 1996). In contrast, outbreaks caused by EHEC are more common, with a number of large foodborne outbreaks being reported in many countries, including Australia (Goldwater and Bettelheim, 1998). In developing countries, the incidence of EHEC infection is reported to be much lower than that of ETEC and EPEC infection (Nataro and Kaper, 1998). EIEC stains have been isolated with low frequency from diarrhoeal cases in both industrialised and less developed countries (Nataro and Levine, 1994). Outbreaks have occurred in hospitals, on a cruise ship, and from contaminated water (Desmarchelier and Fegan, 2003). ETEC stains are a major cause of diarrhoea in infants and young children in developing countries, particularly in the tropics, and are a leading cause of travellers’ diarrhoea (Gross and Rowe, 1985; Doyle and Padhye, 1989; Nataro and Levine, 1994). Although uncommon, a number of foodborne outbreaks due to ETEC have occurred internationally (Olsvik et al., 1991). Mead et al. (1999) estimated that ETEC infection is responsible for approximately 0.4% of foodborne illnesses in the US. In 1983 a multi-state ETEC outbreak occurred in the US that was associated with consumption of imported Brie and Camembert cheese (Anon, 1984; MacDonald et al., 1985). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 67 EPEC stains have caused infantile diarrhoea in hospitals and nurseries in the UK and the US (Robins-Brown 1987; Nataro and Levine, 1994). In developing countries, EPEC stains are still responsible for a high incidence of sporadic infant diarrhoea. Limited information is available on foodborne outbreaks associated with EPEC. An outbreak of EPEC (serotype O111) occurred amongst people on a coach trip to France, although no specific food was identified. The infection was believed to have been the result of consuming food at a restaurant in northern France (Wight et al., 1997). In the US, consumption of undercooked hamburger meat has been an important cause of EHEC outbreaks (Nataro and Kaper 1998). Since its identification as a human pathogen in 1982, and implication in a number of outbreaks in the US, E. coli O157:H7 has become identified as the most predominant cause of EHEC related disease (FAO/WHO, 2000). It is estimated that 85% of EHEC infections in the US are foodborne (Mead et al., 1999). A large multi-state E. coli O157:H7 outbreak involving consumption of contaminated hamburgers occurred in December 1992 – January 1993 with 732 cases identified, of which 195 were hospitalised and 4 died (Nataro and Kaper 1998). Foodborne outbreaks of E. coli O157:H7 have also been associated with consumption of contaminated fresh produce. In the US, outbreaks occurred in 1995 and 1996 (70 and 49 cases respectively), which were traced to consumption of lettuce (Tauxe, 1997). Studies have shown that E. coli O157:H7 can be transmitted to lettuce plant tissue from soil contaminated with manure and contaminated irrigation water (Solomon et al., 2002). Another large E. coli O157:H7 outbreak occurred in the US in 1996 which was linked to apple juice. Although the low pH of fruit juices will generally not allow the survival and growth of many Enterobacteriaceae, some strains of E. coli O157:H7 may survive due to their high acid-tolerance. In 2002, an outbreak of E. coli O157:H7 in Canada was attributed to the consumption of unpasteurised Gouda cheese (Honish et al., 2005). Over 200 non-O157 STEC serotypes have been isolated from humans, with the World Health Organisation identifying O26, O103, O111 and O145 as the most important foodborne nonO157 serogroups worldwide (WHO, 1998). STEC has been a notifiable disease in most Australia States and Territories since August 1998 (Roche et al., 2001). During the period of 2001 – 2005, the notification rate for STEC (excluding HUS cases) in Australia has been 0.2 – 0.3 cases per 100,000 population per annum (Ashbolt et al., 2002; OzFoodNet, 2003; OzFoodNet, 2004; OzFoodNet, 2005). E. coli O157 has been the most commonly reported serotype. Significant variations in notifications exist between states and territories, and part of this variation is likely to be a result of different practices employed by pathology laboratories when screening faecal samples for toxin producing E. coli (OzFoodNet, 2003). A large EHEC outbreak occurred in South Australia during 1995, which resulted in approximately 200 cases of illness. Twenty-two people aged between 4 months and 12 years developed HUS and were hospitalised and a 4 year old child died. Investigations of the outbreak identified EHEC strain O111:NM (or strain O111:H-, NM for non-motile) as the principal cause of the outbreak. A locally produced uncooked, fermented mettwurst was identified as the vehicle for the pathogen. The product was found to contain a variety of EHEC strains in addition to O111 (Paton and Paton, 1998). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 68 Occurrence in food Humans appear to be the primary reservoir of EIEC, ETEC and EPEC organisms (Desmarchelier and Fegan, 2003). Therefore, contamination of food with these organisms is often due to human faecal contamination, either directly from an infected food handler or indirectly via contaminated water. Very little information is available on the occurrence of these organisms in food. The detection of these organisms in food is difficult, requiring sophisticated methodology and therefore food is not routinely screened for these organisms. In general, EPEC and ETEC organisms are more commonly isolated in foods from developing countries and their presence is associated with poor hygiene (Desmarchelier and Fegan, 2003). EPEC has been isolated from milk products in Iraq as well as from a variety of raw and cooked food in Malaysia (Abbar and Kaddar, 1991; Norazah et al., 1998). In Brazil, EPEC has been isolated from 21.1% of soft cheeses sampled (n=45) and has frequently been isolated from pasteurised milk (da Silva et al., 2001; Araújo et al., 2002). EIEC has only sporadically been isolated from foods (Olsvik et al., 1991). In addition to being a major cause of infantile diarrhoea in developing countries, ETEC organisms are a leading cause of traveller’s diarrhoea, which has been linked to the consumption of contaminated food and water (Nataro and Kaper 1998). ETEC has been isolated from Brazilian fish and shrimp which were harvested from waters contaminated with raw sewage (Teophilo et al., 2002). ETEC has also been detected in sauces at Mexican-style restaurants, and in chilli sauce sold by street vendors in Mexico (Adachi et al., 2002; EstradaGarcia et al., 2002). In general, these sauces had been prepared and handled under poor hygienic conditions. The major reservoir of EHEC organisms appears to be the intestinal tract of ruminants, in particular cattle and sheep (Desmarchelier and Fegan, 2003). E. coli O157:H7 and other EHEC species have been isolated from both healthy and diarrhoeic animals, and individual animals can carry more than one serotype (Anon, 1998). Foods derived from these animals may become contaminated via exposure to faecal material during processing. Prevalence of STEC in raw milk has been determined in a limited number of studies. Caution must be exercised when comparing results between independent studies due to differences in sample size, stage of production where the samples were taken and different methodologies used to isolate the organisms. E. coli O157:H7 is the most widely studied EHEC serovar due to it being associated with a large number of outbreaks worldwide. In general, prevalence of STEC in raw milk is low. Adequate pasteurisation will ensure that STEC is inactivated. Very little information is available of the prevalence of EHEC organisms in food in Australia. Of the limited studies undertaken, the prevalence of E. coli O157:H7 in beef and sheep meat appears to be low, however, the prevalence of non-O157:H7 EHEC serotypes is unknown (Vanderlinde et al., 1998; Vanderlinde et al., 1999; Phillips et al., 2001a; Phillips et al., 2001b). Virulence and infectivity Clinical, pathological and epidemiological characteristics of disease caused by pathogenic E. coli vary between pathotypes and are discussed below. EPEC have technically been defined as “diarrhoeagenic E. coli belonging to serogroups epidemiologically incriminated as pathogens but whose pathogenic mechanisms have not been proven to be related either to heat-labile enterotoxins or heat-stable enterotoxins or to MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 69 Shigella-like invasiveness” (Edelman and Levine, 1983). EPEC cause characteristic attaching and effacing lesions in the intestine, similar to those produced by EHEC, but do not produce Shiga toxins. Attachment to the intestinal wall is mediated by a plasmid-encoded outer membrane protein called the EPEC Adherence Factor in type I EPEC. However, pathogenicity is not strictly correlated to the presence of the EPEC Adherence Factor, indicating that other virulence factors are involved (ICMSF, 1996). ETEC that survive passage through the stomach adhere to mucosal cells of the proximal small intestine and produce a heat-labile toxin and/or a heat-stable toxin. The heat-labile toxins are similar in structure and mode of action to cholera toxin, interfering with water and electrolyte movement across the intestinal epithelium (Desmarchelier and Fegan, 2003). If the volume of accumulated fluid exceeds the normal absorptive capacity of the large intestine, the excess is evacuated as watery diarrhoea. EAEC strains are defined as E. coli strains that do not secrete heat-labile or heat-stable toxins. These strains adhere to cultured human epithelial cells in a characteristic aggregative or “stacked-brick” pattern (Yatsuyanagi et al., 2002). The mechanisms causing enteric disease are not fully understood, however EAEC have been associated with persistent diarrhoea, primarily in infants and children (Desmarchelier and Fegan, 2003). Following ingestion, EIEC invade epithelial cells of the distal ileum and colon. The bacteria multiply within the cytoplasm of the cells, causing cell destruction and ulceration. Pathogenicity is associated with a plasmid-encoded type III secretory apparatus and other plasmid-encoded virulence factors (Desmarchelier and Fegan, 2003). The Shiga toxins (Stx1 and Stx2) of EHEC are closely related, or identical, to the toxins produced by Shigella dysenteriae. Additional virulence factors allow the organism to attach tightly to intestinal epithelial cells, causing what is commonly referred to as attaching-andeffacing lesions. Dose response EPEC: It is thought that only a few EPEC cells are necessary to cause illness in children (FDA 2003). Volunteer studies in adults demonstrated that illness could be caused by ingesting 106 – 1010 cells with sodium bicarbonate to neutralise stomach acidity (Doyle and Padhye, 1989). ETEC: Volunteer studies have shown that 108 – 1010 cells of ETEC are necessary for illness in adults (DuPont et al., 1971) although the infective dose is probably less for infants and children (FDA, 2003). EIEC: Volunteer studies have shown that 108 EIEC cells are necessary to cause illness in adults, with the infectious dose reduced to 106 when ingested with sodium bicarbonate (DuPont et al., 1971). However, the US Food and Drug Administration suggest that as few as 10 cells may be needed to cause illness in adults, based on the organisms similarity with Shigella (FDA 2003). The dose-response relationship for EHEC is complicated by the large number of serotypes and the association of EHEC with a variety of foods. Haas et al. (2000) developed a doseresponse relationship for E. coli O157:H7 based on data from a prior animal study undertaken MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 70 by Pai et al. (1997) which involved oral administration of bacterial suspension to infant rabbits. The model was validated by comparison with two well-documented human outbreaks, one foodborne and the other waterborne. The model estimated that the dose required to result in 50% of the exposed population to become ill was 5 × 105 organisms. The corresponding probability of illness for the ingestion of 100 organisms was 2.6 × 10-4. Dose-response relationships for E. coli O111 and O55 have been developed from human feeding trial data (Haas et al., 2000). The relationship estimated a dose required for 50% of the exposed population to become ill was 2.55 × 106 and the probability of illness for ingestion of 100 organisms was 3.5 × 10-4. Investigations of other known outbreaks of foodborne illness due to E. coli O157:H7 and systematic studies aimed at quantifying the dose–response relationship suggest as few as 1 – 700 EHEC organisms can cause human illness (FDA 2003). Host susceptibility A variety of host factors may be important in the pathogenesis of specific E. coli serotypes. In general, the young and the elderly appear to be more susceptible to pathogenic E. coli infection. Epidemiological studies have identified that children are at higher risk of developing post-diarrhoeal HUS than other age groups (Cummings et al., 2002). References Abbar, F. and Kaddar, H.K. (1991) Bacteriological studies on Iraqi milk products. J Appl Bacteriol. 71(6):497500. Adachi, J.A., Mathewson, J.J., Jiang, Z.D., Ericsson, C.D. and DuPont, H.L. (2002) Enteric pathogens in Mexican sauces of popular restaurants in Guadalajara, Mexico, and Houston, Texas. Ann.Intern.Med 136(12):884-887. Anon. (1984) Update: gastrointestinal illness associated with imported semi-soft cheese. Morbidity and Mortality Weekly Review 33(2):22. Anon (1998) Zoonotic non-O157 shiga toxin-producion Escherichia coli (STEC). Report of a WHO Scientific Working Group meeting. World Health Organization, Geneva, Switzerland. Araújo, V.S., Pagliares, V.A., Queiroz, M.L.P. and Freitas-Almeida, A.C. (2002) Occurrence of Staphylococcus and enteropathogens in soft cheese commercialized in the city of Rio de Janeiro, Brazil. Journal of Applied Microbiology 92(6):1172-1177. Arnold, K.W. and Kaspar, C.W. (1995) Starvation- and stationary-phase-induced acid tolerance in Escherichia coli O157:H7. Appl Environ.Microbiol 61(5):2037-2039. Ashbolt, R., Givney, R., Gregory, J.E., Hall, G., Hundy, R., Kirk, M., McKay, I., Meuleners, L., Millard, G., Raupach, J., Roche, P., Prasopa-Plaizier, N., Sama, M.K., Stafford, R., Tomaska, N., Unicomb, L. and Williams, C. (2002) Enhancing foodborne disease surveillance across Australia in 2001: the OzFoodNet Working Group. Commun Dis Intell. 26(3):375-406. Bell, C. and Kyriakides, A. (1998) E. coli: a practical approach to the organism and its control in foods. Blackwell Science, Oxford, UK. Benjamin, M.M. and Datta, A.R. (1995) Acid tolerance of enterohemorrhagic Escherichia coli. Applied and Environmental Microbiology 61(4):1669-1672. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 71 Codex (2002) Risk profile for enterohemorragic E. coli including the identification of commodities of concern, including sprouts, ground beef and pork (CX/FH 03/5-Add .4). Codex Alimentarius Commission, Rome, Italy. ftp://ftp.fao.org/codex/ccfh35/fh0305de.pdf. Cummings, K.C., Mohle-Boetani, J.C., Werner, S.B. and Vugia, D.J. (2002) Population-based trends in pediatric hemolytic uremic syndrome in California, 1994-1999: substantial underreporting and public health implications. Am J Epidemiol 155(10):941-948. da Silva, Z.N., Cunha, A.S., Lins, M.C., de AM, C.L., Almeida, A.C. and Queiroz, M.L. (2001) Isolation and serological identification of enteropathogenic Escherichia coli in pasteurized milk in Brazil. Rev Saude Publica 35(4):375-379. Dalton, C.B., Gregory, J., Kirk, M.D., Stafford, R.J., Givney, R., Kraa, E. and Gould, D. (2004) Foodborne disease outbreaks in Australia, 1995 to 2000. Commun Dis Intell. 28(2):211-224. Desmarchelier, P.M. and Fegan, N. (2003) Enteropathogenic Escherichia coli. In: Hocking, A.D. eds. Foodborne microorganisms of public health significance. Sixth ed, Australian Institute of Food Science and Technology (NSW Branch), Sydney, Australia, pp267-310. Doyle, M.P. and Padhye, V.V. (1989) Escherichia coli. In: Doyle, M.P. eds. Foodborne bacterial pathogens. Marcel Dekker Inc, New York, USA, pp235-281. Doyle, M.P. and Schoeni, J.L. (1984) Survival and growth characteristics of Escherichia coli associated with hemorrhagic colitis. Appl Environ.Microbiol 48(4):855-856. Doyle, M.P., Zhao, T., Meng, J. and Zhao, S. (1997) Escherichia coli O157H:7. In: Doyle, M.P., Beuchat, L.R., and Montville, T.J. eds. Food Microbiology: Fundamentals and frontiers. ASM Press, Washington, USA, pp171-191. DuPont, H.L., Formal, S.B., Hornick, R.B., Snyder, M.J., Libonati, J.P., Sheahan, D.G., LaBrec, E.H. and Kalas, J.P. (1971) Pathogenesis of Escherichia coli diarrhea. N Engl.J Med 285(1):1-9. Edelman, R. and Levine, M.M. (1983) From the National Institute of Allergy and Infectious Diseases. Summary of a workshop on enteropathogenic Escherichia coli. J.Infect.Dis. 147(6):1108-1118. Estrada-Garcia, T., Cerna, J.F., Thompson, M.R. and Lopez-Saucedo, C. (2002) Faecal contamination and enterotoxigenic Escherichia coli in street-vended chili sauces in Mexico and its public health relevance. Epidemiol Infect 129(1):223-226. European Commission (2000) Opinion of the scientific committee on veterinary measures relating to public health on food-borne zoonoses. Health and Consumer Protection Directorate-General, European Commission, Brussels, Belgium. FAO/WHO (2000) Joint FAO/WHO expert consultation on risk assessment of microbiological hazards in foods. FDA (2003) The bad bug book (Foodborne Pathogenic Microorganisms and Natural Toxins Handbook ). US Food and Drug Administration. http://www.cfsan.fda.gov/~mow/intro.html. Accessed on 8 August 2004. Goldwater, P.N. and Bettelheim, K.A. (1998) New perspectives on the role of Escherichia coli O157:H7 and other enterohaemorrhagic E. coli serotypes in human disease. J.Med.Microbiol. 47(12):1039-1045. Gross, R.J. and Rowe, B. (1985) Escherichia coli diarrhoea. J Hyg.(Lond) 95(3):531-550. Haas, C.N., Thayyar-Madabusi, A., Rose, J.B. and Gerba, C.P. (2000) Development of a dose-response relationship for Escherichia coli O157:H7. Int J Food Microbiol 56(2-3):153-159. Honish, L., Predy, G., Hislop, N., Chui, L., Kowalewska-Grochowska, K., Trottier, L., Kreplin, C. and Zazulak, I. (2005) An outbreak of E. coli O157:H7 hemorrhagic colitis associated with unpasteurized gouda cheese. Can.J Public Health 96(3):182-184. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 72 ICMSF. (1996) Intestinally pathogenic Escherichia coli. In: Roberts, T.A., Baird-Parker, A.C., and Tompkin, R.B. eds. Microorganisms in foods 5: Microbiological specifications of food pathogens. First ed, Blackie Academic & Professional, London, UK, pp217-264. Kirk, R. and Rowe, M. (1999) Pathogenic E. coli and milk. Milk Industry Int 101((Technical and Research Supplement MII)):4. Lake, R., Hudson, A. and Cressey, P. (2003) Risk Profile: Shiga-like toxin producing Escherichia coli in uncooked comminuted fermented meat products. Institute of Environmental Science and Research Limited, Christchurch, New Zealand. MacDonald, K.L., Eidson, M., Strohmeyer, C., Levy, M.E., Wells, J.G., Puhr, N.D., Wachsmuth, K., Hargrett, N.T. and Cohen, M.L. (1985) A multistate outbreak of gastrointestinal illness caused by enterotoxigenic Escherichia coli in imported semisoft cheese. J Infect Dis 151(4):716-720. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M. and Tauxe, R.V. (1999) Food-related illness and death in the United States. Emerg.Infect Dis 5(5):607-625. Nataro, J.P. and Kaper, J.B. (1998) Diarrheagenic Escherichia coli. Clin Microbiol Rev 11(1):142-201. Nataro, J.P. and Levine, M.M. (1994) Escherichia coli diseases in humans. In: Gyles, C.L. eds. Escherichia coli in domestic animals and humans. CAB International, Wallingford, UK, pp285-333. Norazah, A., Rahizan, I., Zainuldin, T., Rohani, M.Y. and Kamel, A.G. (1998) Enteropathogenic Escherichia coli in raw and cooked food. Southeast Asian J Trop.Med Public Health 29(1):91-93. Olsvik, Ø., Wasteson, Y., Lund, A. and Hornes, E. (1991) Pathogenic Escherichia coli found in food. Int J Food Microbiol 12(1):103-113. OzFoodNet (2003) Foodborne disease in Australia: incidence, notification and outbreaks. Annual Report of the OzFoodNet Network, 2002. Communicable Diseases Intelligence 27(2):http://www.cda.gov.au/pubs/cdi/2003/cdi2702/htm/cdi2702g.htm. OzFoodNet (2004) Foodborne disease investigations across Australia: Annual report of the OzFoodNet network, 2003. Commun Dis Intell. 28(3):359-389. OzFoodNet (2005) Foodborne outbreaks associated with dairy products — Analysis of OzFoodNet data, 1995 – June 2004. Unpublished. Pai, M., Kang, G., Ramakrishna, B.S., Venkataraman, A. and Muliyil, J. (1997) An epidemic of diarrhoea in south India caused by enteroaggregative Escherichia coli. Indian J.Med.Res. 106:7-12. Paton, J.C. and Paton, A.W. (1998) Pathogenesis and diagnosis of Shiga toxin-producing Escherichia coli infections. Clin Microbiol Rev 11(3):450-479. Phillips, D., Sumner, J., Alexander, J.F. and Dutton, K.M. (2001a) Microbiological quality of Australian beef. J Food Prot. 64(5):692-696. Phillips, D., Sumner, J., Alexander, J.F. and Dutton, K.M. (2001b) Microbiological quality of Australian sheep meat. J Food Prot. 64(5):697-700. Robins-Brown, R.M. (1987) Traditional enteropahogenic Escherichia coli of infantile diarrhea. Rev.Infect.Dis. 9:28-53. Roche, P., Spencer, J., Lin, M., Gidding, H., Kirk, M., Eyeson-Annan, M., Milton, A., Witteveen, D. and Merianos, A. (2001) Australia's notifiable diseases status, 1999: annual report of the National Notifiable Diseases Surveillance System. Commun Dis Intell. 25(4):190-245. Solomon, E.B., Potenski, C.J. and Matthews, K.R. (2002) Effect of irrigation method on transmission to and persistence of Escherichia coli O157:H7 on lettuce. J Food Prot. 65(4):673-676. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 73 Tauxe, R.V. (1997) Emerging foodborne diseases: an evolving public health challenge. Emerg.Infect Dis 3(4):425-434. Teophilo, G.N., dos, F., V, dos Prazeres, R.D. and Menezes, F.G. (2002) Escherichia coli isolated from seafood: toxicity and plasmid profiles. Int Microbiol 5(1):11-14. Vanderlinde, P.B., Shay, B. and Murray, J. (1998) Microbiological quality of Australian beef carcass meat and frozen bulk packed beef. Journal of Food Protection 61(4):437-443. Vanderlinde, P.B., Shay, B. and Murray, J. (1999) Microbiological status of Australian sheep meat. Journal of Food Protection 62(4):380-385. WHO (1998) Zoonotic non-O157 shiga toxin-producing Escherichia coli (STEC). Preport of a WHO Scientific Working Group Meeting. Berlin, Germany, 26-26 June 1998. World Health Organization, Geneva. Wight, J.P., Rhodes, P., Chapman, P.A., Lee, S.M. and Finner, P. (1997) Outbreaks of food poisoning in adults due to Escherichia coli O111 and campylobacter associated with coach trips to northern France. Epidemiol Infect 119(1):9-14. Yatsuyanagi, J., Saito, S., Sato, H., Miyajima, Y., Amano, K. and Enomoto, K. (2002) Characterization of enteropathogenic and enteroaggregative Escherichia coli isolated from diarrheal outbreaks. J Clin Microbiol 40(1):294-297. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 74 3. Listeria monocytogenes Listeria monocytogenes is a Gram-positive, non-spore forming rod-shaped bacteria that may be isolated from a variety of sources including soil, silage, sewage, food-processing environments, raw meats and the faeces of healthy humans and animals (FDA, 2003). L. monocytogenes belongs to the genus Listeria along with L. innocua, L. welshimeri, L. selligeri, L. ivanovii and L. grayi. Thirteen serotypes are associated with L. monocytogenes (1/2a, 1/2b, 1/2c, 3a, 3b, 3c, 4a, 4ab, 4b, 4c, 4d, 4e, 7). Growth characteristics Growth of L. monocytogenes in foods is influenced by a variety of factors, including the nature and concentration of essential nutrients, pH, temperature, water activity, the presence of food additives that could enhance or inhibit growth and presence of other microbial flora (Lovett et al., 1987). Under conditions outside the growth range, the bacteria may survive and growth may recommence once suitable conditions are encountered. Temperatures of >50ºC are lethal to L. monocytogenes. When in a suitable medium, L. monocytogenes can grow between ~0 - 45°C. Although L. monocytogenes does not grow below –1.5ºC, it can readily survive at much lower temperatures. Nonetheless, freezing and frozen storage will cause a limited reduction in the viable population of L. monocytogenes. Optimal conditions for growth are between 30 - 37ºC (Ryser and Marth, 1999). L. monocytogenes will grow in a broad pH range with the upper limit being approximately 9.3 and the lower limit being 4.6-5.0 (ICMSF, 1996). Although growth at pH <4.3 has not yet been documented, L. monocytogenes appears to be relatively acid tolerant. It has been suggested that food fermentations, which involve a gradual lowering of pH, could lead to acid adaptation of L. monocytogenes. Like many bacterial species, L. monocytogenes grows optimally at a of approximately 0.97. However, when compared with most foodborne pathogens, the bacterium has the unique ability to multiply at water activity values as low as 0.90. While it does not appear to be able to grow below 0.90, the bacterium can survive for extended periods at lower values (Ryser and Marth, 1999). L. monocytogenes is reasonably tolerant to salt and can grow in NaCl concentrations up to 10%. Extended survival occurs at a wide range of salt concentrations and L. monocytogenes has survived for up to eight weeks in a concentration of 20% NaCl (Sutherland et al., 2003). Survival in the presence of salt varies with storage temperature and studies have indicated that survival of L. monocytogenes in concentrated salt solutions can be increased dramatically by lowering the incubation temperature (Ryser and Marth, 1999). L. monocytogenes grows well under both aerobic and anaerobic conditions (Ryser and Marth, 1999; Sutherland et al., 2003). The listericidal effect of preservatives is strongly influenced by the interactive effects of temperature, pH, type of acidulant, salt content, water activity, and type and concentration of food additives present in the food. For example, the ability of potassium sorbate to prevent growth of L. monocytogenes is related to temperature and pH. The lower the storage temperature and pH of the medium, the greater the effectiveness of sorbates against L. monocytogenes. Sodium benzoate is more inhibitory to L. monocytogenes than is either potassium sorbate or sodium propionate. Inhibition and inactivation of L. monocytogenes in the presence of sodium benzoate is affected by temperature (more rapid at higher than lower MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 75 incubation temperatures), concentration of benzoic acid (more rapid at higher than lower concentrations) and pH (more rapid at lower rather than higher pH values) as well as the type of acid used to adjust the growth medium (Ryser and Marth, 1999). Pathology of illness There are two main forms of illness associated with L. monocytogenes infection: listerial gastroenteritis, where usually only mild symptoms are reported, and invasive listeriosis, where the bacteria penetrate the gastrointestinal tract and invade normally sterile sites within the body (FDA 2003). Symptoms of the mild form of L. monocytogenes infection are primarily those generally associated with gastrointestinal illness: chills, diarrhoea, headache, abdominal pain and cramps, nausea, vomiting, fatigue, and myalgia (FDA, 2003). The onset of illness is usually greater than 12 hours. Invasive listeriosis is clinically defined when the organism is isolated from blood, cerebrospinal fluid or an otherwise normally sterile site (e.g. placenta, foetus). The manifestations include septicaemia, meningitis (or meningoencephalitis), encephalitis, and intrauterine or cervical infections in pregnant women, which may result in spontaneous abortion in the second or third trimester, or stillbirth (FDA, 2003). The onset of these manifestations is usually preceded by influenza-like symptoms including persistent fever. Gastrointestinal symptoms such as nausea, vomiting and diarrhoea may also precede the serious forms of listeriosis. Listeriosis typically has a 2 - 3 week incubation time, but onset time may extend to 3 months (FDA/Centre for Food Safety and Applied Nutrition, 2003). It is estimated that approximately 2 – 6% of the healthy human population harbour L. monocytogenes in their intestinal tract, which suggests that people are frequently exposed to L. monocytogenes (Farber and Peterkin, 1991; Rocourt and Bille, 1997). This may also suggest that most people have a tolerance to infection by L. monocytogenes, and given the relatively low number of reported cases, exposure rarely leads to serious illness in healthy individuals (Hitchins, 1996; Marth, 1988). Mode of transmission Foodborne exposure is the primary route of transmission for listeriosis, however listeriosis can be transmitted vertically (i.e. mother to child), zoonotically and through hospital acquired infections (Ryser and Marth, 1999; Bell and Kyriakides, 2005). Incidence of illness Most cases of listeriosis are sporadic. The number of reported cases of invasive listeriosis in Australia between 2001 - 2004 varied between 61 – 72 cases (Ashbolt et al., 2002; Anon, 2002; Anon, 2003; Anon, 2004b), which equates to approximately 3 – 4 cases per million population per annum. In Australia, the exact mortality rate is not known, although the data available would suggest a rate of approximately 23%. The case fatality rate in New Zealand is approximately 17% (Anon, 2004a). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 76 The estimated incidence of invasive listeriosis in European countries has been reported to between 0.3 - 7.5 cases per million of the general population per annum (European Commission, 2003). In France, the estimated incidence is sixteen cases per million (general population) per annum (Bille, 1990; ICMSF, 1996). The annual incidence of listeriosis in the United States has been estimated to range from 3.4 per million (CDC, 2002) to 4.4 per million (Tappero et al., 1995). Of all foodborne pathogens, L. monocytogenes results in the highest hospitalisation rate in the US, with fatality rates of 20 - 30% being common (WHO/FAO, 2004). Outbreaks of invasive listeriosis have been linked to Hispanic-style soft cheeses; soft, semi-soft and mould-ripened cheeses; hot dogs; pork tongue jelly; processed meats; pate; salami; pasteurised chocolate flavoured milk; pasteurised and unpasteurised milk; butter; cooked shrimp; smoked salmon; maize and rice salad; maize and tuna salad; potato salad; raw vegetables; and coleslaw (FDA 2003). In addition, sporadic cases have been linked to the consumption of raw milk; unpasteurised ice cream; ricotta cheese; goat, sheep and feta cheeses; soft, semi-soft and mould-ripened cheeses; Hispanic-style cheese; salami; hot dogs; salted mushrooms; smoked cod roe; smoked mussels; undercooked fish; pickled olives; raw vegetables; and coleslaw (WHO/FAO, 2004). Occurrence in foods L. monocytogenes has been found in foods such as milk, dairy products (particularly softripened cheeses), meat, poultry, seafood and vegetables. The worldwide prevalence of L. monocytogenes in raw milk is estimated to be around 3-4% (Hayes et al., 1986; Lovett et al., 1987; Doores and Amelang, 1988). In Australian surveys on soft and surface ripened cheeses and ice-cream, L. monocytogenes has been isolated from 2% of locally produced cheese samples and 6% of ice-cream samples (Sutherland et al., 2003). For imported cheeses, camembert and blue vein, 7% were positive for L. monocytogenes (Sutherland et al., 2003). For European soft and surface-ripened cheeses, 25% have been found to be positive for L. monocytogenes (Terplan, 1988). Meat products from which L. monocytogenes has been isolated include beef, lamb, pork, minced meat products, sausages, salami, ham, mettwurst, pate, frankfurters and vacuumed packed meat, chicken products, and processed seafood (Farber and Peterkin 1991; Cox et al., 1999; Ojeniyi et al., 2000). Additionally vegetable products have also been shown to be contaminated (Heisick et al., 1989; Brackett, 1999). Virulence and infectivity of L. monocytogenes When ingested, L. monocytogenes penetrates the intestinal tissue and is taken up by macrophages and non-phagocytic cells in the host. L. monocytogenes is disseminated throughout the host via blood or lymphatic circulation to various tissues. Its presence intra-cellularly in phagocytic cells permits access to the brain and probably transplacental migration to the foetus in pregnant women. The pathogenesis of L. monocytogenes relies on its ability to survive and multiply in phagocytic host cells. Not all strains appear to be equally virulent. The 4b and occasionally 1/2a and 1/2b serovars account for most cases of human listeriosis (ICMSF, 1996). The virulence of L. monocytogenes is increased when the bacterium is grown at low rather than high temperatures. The possibility exists that cold storage may enhance the virulence of some L. monocytogenes strains isolated from refrigerated foods (Ryser and Marth, 1999). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 77 Dose response Cases of non-invasive listeriosis (also referred to as febrile listerial gastroenteritis) have been observed during outbreaks, involving symptoms such as diarrhoea, fever, headache and myalgia, generally following a short incubation period (WHO/FAO, 2004). Insufficient quantitative data is available to develop a dose-response model for this milder form of listeriosis, however, outbreak situations have generally involved the ingestion of high doses of L. monocytogenes. The dose-response relationship for invasive listeriosis is highly dependent on a number of factors, such as the virulence characteristics of the organism, the number of cells ingested, the general health and immune status of the host, and the attributes of the food matrix that may alter the microbial or host status. WHO/FAO (2004) and FDA/FSIS (2003) developed separate dose-response models for both healthy and susceptible populations by combining data from surrogate animal models with epidemiological data. The Exponential dose-response model was used for both populations. This dose-response model has a single parameter, the r-value. The r-value is the probability that a person will become ill from the consumption of a single L. monocytogenes cell. For the healthy population (classified as “intermediate-age”) the median r-value was estimated to be 2.37 x 10-14. For more susceptible populations the median r-value was estimated to be 1.06 x 10-12. A more recent assessment of US epidemiological data on invasive Listeriosis in susceptible sub-populations which included genetic information regarding different L. monocytogenes strains (lineages), determined average r-values of 1.31 x 10-8 for lineage I and 5.01 x 10-11 for lineage II (Chen et al., 2006). Further analysis of the epidemiological data by the L. monocytogenes ribotype found r-values as small as 6.29 x 10-3. These results suggest that there are large differences in virulence between L. monocytogenes strains. The infectious dose is unknown but it is believed to vary depending on the strain and susceptibility of the individual. There is a lack of information concerning the minimal infectious dose, although it is generally thought to be relatively high (>100 viable cells) (ICMSF, 1996). From cases contracted via raw or inadequately pasteurised milk, it is assumed that for susceptible individuals, ingestion of fewer than 1,000 organisms may cause disease (FDA/FSIS, 2003). It is thought the consumption of food with exceptionally high levels of L. monocytogenes (>107/g) is required to cause the mild gastrointestinal form of illness in healthy persons (Sutherland et al., 2003). Host factors Specific sub-populations at risk for invasive listeriosis include pregnant women and their foetuses, neonates, the elderly and persons with a compromised immune system, whose resistance to infection is lowered (e.g. transplant patients, patients on corticosteroid treatments, AIDS patients and alcoholics). Less frequently reported diabetic, cirrhotic, asthmatic and ulcerative colitis patients are also at a higher risk (FDA 2003). Another physiological parameter thought to be relevant to susceptibility is a reduced level of gastric acidity (WHO/FAO, 2004). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 78 Food matrix To date, the properties of the food vehicle have been viewed as having little effect on the infective dose of L. monocytogenes. However, it is possible that food vehicles with high buffering capacity may protect the bacteria from inactivation by the pH of gastric acids in the stomach. In general, there are insufficient data available as to whether the food matrix affects the dose-response curve for L. monocytogenes (WHO/FAO, 2004). References Anon. (2002) OzFoodNet: enhancing foodborne disease surveillance across Australia: quarterly report, April to June 2002. Commun Dis Intell. 26(4):546-551. Anon. (2003) Foodborne disease in Australia: incidence, notifications and outbreaks. Annual report of the OzFoodNet network, 2002. Commun Dis Intell. 27(2):209-243. Anon (2004a) Notifiable and other diseases in New Zealand: Annual report, 2004. Ministry of Health, Wellington, New Zealand. Anon. (2004b) OzFoodNet: enhancing foodborne disease surveillance across Australia: quarterly report, 1 October to 31 December 2003. Commun Dis Intell. 28(1):86-89. Ashbolt, R., Givney, R., Gregory, J.E., Hall, G., Hundy, R., Kirk, M., McKay, I., Meuleners, L., Millard, G., Raupach, J., Roche, P., Prasopa-Plaizier, N., Sama, M.K., Stafford, R., Tomaska, N., Unicomb, L. and Williams, C. (2002) Enhancing foodborne disease surveillance across Australia in 2001: the OzFoodNet Working Group. Commun Dis Intell. 26(3):375-406. Bell, C. and Kyriakides, A. (2005) Listeria. Second ed, Blackwell Publishing, UK. Bille, J. (1990) Anatomy of a foodborne listeriosis outbreak. In: Miller, B., Smith, J.L., and Somkuti, G. eds. Foodborne Listeriosis. Elsevier, London, pp39-46. Brackett, R.E. (1999) Incidence and behaviour of Listeria monocytogenes in products of plant origin. In: Ryser, E.T. and Marth, E.H. eds. Listeria, Listeriosis and food safety. Marcel Dekker, New York, pp631-655. CDC (2002) Foodborne Outbreak Response Surveillance Unit. http://www2a.cdc.gov/ncidod/foodborne/fbsearch.asp. Accessed on 10 May 5 A.D. Cox, N.A., Bailey, J.S. and Ryser, E.T. (1999) Incidence and behaviour of Listeria monocytogenes in poultry and egg products. In: Ryser, E.T. and Marth, E.H. eds. Marcel Dekker, New York, pp565-600. Doores, S. and Amelang, J. (1988) Incidence of Listeria monocytogenes in the raw milk supply in Pennsylvania. (1) IFT Ann. Meeting, 19-22 June., FT Chicago, New Orleans, pp157. European Commission (2003) Opinion of the Scientific Committee of Veterinary Measures Relating to Public Health on Staphylococcal Enterotoxins in Milk Products, particularly cheeses. http://europa.eu.int/comm/food/fs/sc/scv/out61_en.pdf. Farber, J.M. and Peterkin, P.I. (1991) Listeria monocytogenes, a food-borne pathogen. Microbiol.Rev. 55(3):476-511. FDA (2003) The bad bug book (Foodborne Pathogenic Microorganisms and Natural Toxins Handbook ). US Food and Drug Administration. http://www.cfsan.fda.gov/~mow/intro.html. Accessed on 8 August 2004. FDA/Centre for Food Safety and Applied Nutrition (2003) Quantitative Assessment of the Relative Risk to Public Health from Foodborne Listeria monocytogenes Among Selected Categories of Ready-to-Eat Foods. http://www.foodsafety.gov/~dms/lmr2-toc.html. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 79 FDA/FSIS. (2003) Quantitative assessment of relative risk to public health from foodborne Listeria monocytogenes among selected categories of ready-to-eat foods. Food and Drug Administration, US Dept of Health and Human Services; Food Safety Inspection Service, United States Department of Agriculture, USA. Hayes, P.S., Feeley, J.C., Graves, L.M., Ajello, G.W. and Fleming, D.W. (1986) Isolation of Listeria monocytogenes from raw milk. Appl.Environ.Microbiol. 51(2):438-440. Heisick, J.E., Wagner, D.E., Nierman, M.L. and Peeler, J.T. (1989) Listeria spp. found on fresh market produce. Appl.Environ.Microbiol. 55(8):1925-1927. Hitchins, A.D. (1996) Assessment of alimentary exposure to Listeria monocytogenes. Int.J Food Microbiol. 30(1-2):71-85. ICMSF. (1996) Microorganisms in Food 5: Microbiological Specifications of Food Pathgens. Blackie Academic and Professional, London. Lovett, J., Francis, D.W. and Hunt, J.M. (1987) Listeria monocytogenes in raw milk; detection, incidence and pathogenicity. J Food Prot 50:188-192. Marth, E.H. (1988) Disease Characteristics of Listeria monocytogenes. Food Technol 42(4):165-168. Ojeniyi, B., Christensen, J. and Bisgaard, M. (2000) Comparative investigations of Listeria monocytogenes isolated from a turkey processing plant, turkey products, and from human cases of listeriosis in Denmark. Epidemiol.Infect. 125(2):303-308. Rocourt, J. and Bille, J. (1997) Foodborne listeriosis. World Health Stat.Q. 50(1-2):67-73. Ryser, E.T. and Marth, E.H. (1999) Listeria, Listeriosis, and Food Safety. Marcel Dekker, Inc., New York. Sutherland, P.S., Miles, D.W. and Laboyrie, D.A. (2003) Listeria monocytogenes. In: Hocking, A.D. eds. Foodborne microorganisms of public health significance. Australian Institute of Food Science and Technology (NSW Branch, Food Microbiology Group), Waterloo DC, NSW, Australia, pp381-443. Tappero, J.W., Schuchat, A., Deaver, K.A., Mascola, L. and Wenger, J.D. (1995) Reduction in the incidence of human listeriosis in the United States. Effectiveness of prevention efforts? The Listeriosis Study Group. JAMA 273(14):1118-1122. Terplan, G. (1988) Listeria in the dairy industry. Proceedings of Gehr's seminar on problems of foodborne Listeriosis, European Symposium. (7 Sept) European Symposium, Wiesbaden, Germany. WHO/FAO (2004) Risk assessment of Listeria monocytogenes in ready-to-eat foods. World Health Organization and Food and Agriculture Organization of the United Nations, Geneva, Switzerland. http://www.fao.org/es/ESN/food/risk_mra_riskassessment_listeria_en.stm. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 80 4. Salmonella spp. Salmonellosis is a leading cause of enteric illness, with symptoms ranging from mild gastroenteritis to systemic illness such as septicaemia and other longer-term conditions. A wide range of foods have been implicated in foodborne salmonellosis. However, as the disease is primarily zoonotic, foods of animal origin have been consistently implicated as the main sources of human salmonellosis (FAO/WHO, 2002). The genus Salmonella is currently divided into two species: Salmonella enterica (comprising six subspecies) and Salmonella bongori (Brenner et al., 2000). The subspecies of most concern in relation to food safety is S. enterica subsp. enterica, as over 99% of human pathogens belong to this subspecies (Bell and Kyriakides, 2002). Over 1,400 S. enterica subsp. enterica serotypes are currently recognised, and all are regarded as capable of causing illness in humans (Brenner et al., 2000). The formal names to describe Salmonella serotypes are rather cumbersome, for example S. enterica subsp. enterica serotype Typhimurium (formerly Salmonella typhimurium). For practical reasons, the shortened versions of these names are commonly used, such as Salmonella Typhimurium. Some Salmonella serotypes are host-adapted to individual animal species. For example S. Typhi and S. Paratyphi are specifically associated with infections leading to severe illness in humans (Bell and Kyriakides, 2002). Growth characteristics Salmonellae have relatively simple nutritional requirements and can survive for long periods of time in foods and other substrates. The rate of growth and extent of survival of the organism in a particular environment is influenced by the simultaneous effect of a number of factors such as temperature, pH, and water activity. Being facultative anaerobic, salmonellae also have the ability to grow in the absence of oxygen. Growth and survival is also influenced by the presence of inhibitors such as nitrite and short-chain fatty acids (Jay et al., 2003). The growth of most salmonellae is substantially reduced at temperatures <15°C and prevented at <7°C. Growth generally does not occur at temperatures >46.2°C. The optimum temperature for growth is 35 – 43°C. Freezing can be detrimental to Salmonella spp. survival, although it does not guarantee destruction of the organism (ICMSF, 1996). There is an initial rapid decrease in the number of viable organisms at temperatures close to freezing point as a result of freezing damage. However, at lower temperatures (-17 to -20°C) there is a significantly less rapid decline in the number of viable organisms. Salmonella spp. have the ability to survive long periods of time at storage temperatures < -20°C (Jay et al., 2003). Heat resistance of Salmonella spp. in foods is dependant on the composition, nature of solutes, pH, and water activity of the food (Jay et al., 2003). In general, heat resistance increases as the water activity of the food decreases. A reduction in pH results in a reduction of heat resistance (ICMSF, 1996). The minimum pH at which Salmonella spp. can grow is dependent on the temperature of incubation, the presence of salt and nitrite and the type of acid present. However, growth can usually occur between pH 3.8 – 9.5 (Jay et al., 2003). The optimum pH range for growth is 7.0 – 7.5. Volatile fatty acids are more bactericidal than acids such as lactic and citric acid. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 81 Water activity has a significant effect on the growth of Salmonella spp., with the lower limit for growth being 0.94 (ICMSF, 1996). Salmonella can survive for long periods of time in foods with a low water activity (such as black pepper, chocolate, gelatine). Exposure to low water activity environments can greatly increase the heat resistance of Salmonella spp. Pathology of illness Outcomes of exposure to Salmonella spp. can range from having no effect, to colonisation of the gastrointestinal tract without symptoms of illness (asymptomatic), or colonisation with the typical symptoms of acute gastroenteritis (FAO/WHO, 2002). Gastroenteritis symptoms may include abdominal pain, nausea, diarrhoea, mild fever, vomiting, headache and/or prostration, with clinical symptoms lasting 2 – 5 days. Most symptoms of salmonellosis are mild, and only a low proportion of cases within the community are reported to public health agencies (Mead et al., 1999). In a small number of cases, Salmonella spp. infection can lead to more severe invasive diseases characterised by septicaemia and sometimes death. In a study of 48,857 patients with gastroenteritis (of which 26,974 were salmonellosis), Helms et al. (2003) found an association with increased short-term (mortality within 30 days of infection) and long-term (mortality within a year of infection) risk of death compared with controls. In cases of acute gastroenteritis, the incubation period is usually 12 - 72 hours (commonly 12 - 36 hours) and is largely dependant on the sensitivity of the host and size of the dose ingested (Hohmann, 2001; FAO/WHO, 2002). Illness is usually self-limiting, with patients fully recovering within one week, although in some severe cases of diarrhoea, significant dehydration can ensue which may require medical intervention such as intravenous fluid replacement. Septicaemia is caused when Salmonella spp. enters the bloodstream, with symptoms including high fever, pain in the thorax, chills, malaise and anorexia (FAO/WHO, 2002). Although uncommon, long-term effects or sequelae may occur including arthritis, appendicitis, cholecystitis, endocarditis, local abscesses, meningitis, osteomyelitis, osteoarthritis, pericarditis, peritonitis, pleurisy, pneumonia and urinary tract infection (ICMSF, 1996). At the onset of illness large numbers of Salmonella spp. are excreted in the faeces. Numbers decrease with time, but the median duration of excretion after acute nontyphoid salmonellosis has been estimated at five weeks, and approximately 1% of patients become chronic carriers (Jay et al., 2003). Due to the general self-limiting nature of the disease, antibiotics are not usually recommended for healthy individuals suffering from mild to moderate Salmonella spp. gastroenteritis (Hohmann 2001). Antibiotics should be used, however, for those who are severely ill and for patients with risk factors for extra intestinal spread of infection, after appropriate blood and faecal cultures are obtained. Of recent concern worldwide is the emergence of multiple antibiotic resistant strains of Salmonella spp., an example being S. Typhimurium definitive phage type 104 (DT104). Multi-resistant S. Typhimurium DT104 is a significant human and animal pathogen, with high morbidity observed in cattle and poultry (Crerar et al., 1999). To date, this organism is not endemic in Australia, although it is a significant health problem in European countries, North America, the Middle East, South Africa and South-East Asia (Jay et al., 2003). S. Typhimurium DT104 constitutes 8 – 9% of human Salmonella spp. isolates in the US. Sporadic human cases are reported in Australia, although these are commonly acquired overseas (Blumer et al., 2003). During 2001 an outbreak of S. Typhimurium DT104 occurred in Victoria and was linked to contaminated imported halva (a sesame seed product). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 82 Mode of transmission Salmonellae are transmitted by the faecal-oral route. Sources of transmission include personto-person, foodborne, waterborne (drinking water and direct contact with faecally contaminated water) and direct contact with infected animals. Incidence and outbreak data Salmonellosis is one of the most commonly reported enteric illnesses worldwide (FAO/WHO, 2002). Approximately 7,000 - 8,000 cases of salmonellosis per annum are formally notified to health authorities in Australia. Taking into account under-reporting it has been estimated (based on published rates of under-reporting) that 80,000 cases of foodborne salmonellosis occur annually (Hall, 2003). The salmonellosis notification rate in Australia for 2002 was 40.3 cases per 100,000 population. This varied from 24.8 cases per 100,000 population in Victoria to 166.7 cases per 100,000 population in the Northern Territory (Anon, 2003). Children less than five years of age have by far the highest notification rate, with a rate of 210.6 cases per 100,000 population reported for 2002 (Yohannes et al., 2004). The higher rate of notified salmonellosis cases in this age group may reflect an increased susceptibility upon first exposure, but may also be a result of other factors such as an increased likelihood of exposure and increased likelihood to seek medical care and be tested. Of the total number of Salmonella serovars reported to Australian health authorities during 2002, S. Typhimurium 135 was the most commonly reported. Distribution of Salmonella serovars varies geographically, with the most commonly reported serovars in Queensland, Tasmania and the Northern Territory being S. Virchow (10%), S. Mississippi (48%) and S. Ball (15%) respectively. Of the other States and Territories, S. Typhimurium was the most commonly reported serovar, representing 34% of cases in the Australian Capital Territory, 28% in New South Wales, 60% in South Australia, 66% in Victoria and 15% in Western Australia. Salmonellosis notifications in Australia fluctuate seasonally, from a low in August September to a peak in January - March, with 36% of salmonellosis cases notified during this period (Yohannes et al., 2004). It has been estimated that in the US (Mead et al., 1999) and England and Wales (Adak et al., 2002), 95% and 91.6%, respectively of salmonellosis cases are foodborne. Other sources of infection may be via contaminated water, person-to-person transmission and direct contact with infected animals. Based on results from national and international epidemiological data (primarily outbreak investigations) a wide range of foods have been implicated in human salmonellosis. Foods of animal origin (e.g. meat, eggs, and dairy) are important sources of human salmonellosis. Following notifications of salmonellosis to Australian health authorities, over 50 epidemiological investigations are initiated each year in an attempt to identify a common source of infection (Anon 2003). It is often difficult, however, to confirm a single food commodity as a source due to the difficulty of investigating commonly consumed foods, conducting trace-back, and lack of systematically collected microbiological data from foods. In a review of reported foodborne disease outbreaks in Australia during 1995 – 2000, meats (in particular poultry meat) were associated with 33% of identified salmonellosis outbreaks (Dalton et al., 2004). A large outbreak (consisting of 502 cases) of S. Typhimurium 135a occurred in 1999 and was associated with consumption of unpasteurised commercial orange juice (Roche et al., 2001). In 2001 a community-wide outbreak of S. Typhimurium 126 MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 83 occurred in South Australia (Ashbolt et al., 2002). A subsequent case-control study associated illness with the consumption of chicken meat. This link was corroborated with microbiological testing of raw poultry, and the likely source of contaminated products was traced to a single poultry processing facility. Occurrence in food The primary reservoir of Salmonella spp. is the intestinal tract of warm and cold-blooded vertebrates. Infected animals shed large numbers in their faeces, and this leads to contamination of the surrounding environment including soil, pasture, streams and lakes. Salmonella spp. have been isolated from a wide range of foods, particularly those of animal origin and those foods that have been subject to faecal contamination (ICMSF, 1996). Raw meat products (in particular poultry) have frequently been associated with the presence of Salmonella spp. Salmonella spp. positive animals at the time of slaughter may have high numbers of organisms in their intestines as well as on external surfaces (faecal contamination of hides, fleece, skin or feathers). Cross contamination during processing may also lead to increased prevalence of Salmonella spp. in finished products (Bryan and Doyle 1995). Pasteurisation of dairy products effectively inactivates Salmonella spp., however contamination of milk has occurred due to improper pasteurisation and/or post-processing contamination (Jay et al., 2003). Virulence and infectivity Once ingested, Salmonella spp. must be able to overcome the low pH of the stomach, adhere to the small intestine epithelial cells and overcome host defence mechanisms to enable infection (Jay et al., 2003). Salmonella spp. possess a number of structural and physiological virulence factors enabling them to cause acute and chronic disease in humans. Virulence of Salmonella spp. varies with the length and structure of the O side chains of lipopolysaccharide molecules at the surface of the cell. Resistance of Salmonella spp. to the lytic action of complement is directly related to the length of the O side chain (Jay et al., 2003). The presence of virulence plasmids has been associated with the ability to spread rapidly after colonisation and overwhelm the host immune response (D'Aoust, 1997). These virulence plasmids are large cytoplasmic DNA structures that replicate independently of the chromosomal DNA. Virulence plasmids are present in a limited number of Salmonella serovars and have been confirmed in S. Typhimurium, S. Dublin, S. Gallinarum, S. Pullorum, S. Enteritidis, S. Choleraesuis and S. Abortusovis. It is notable, however, that virulence plasmids are absent from S. Typhi, which is host-adapted and highly infectious. Once attached to small intestine epithelial cells, the organism is drawn into the host cell in a vesicle (endosome) where it can multiply in the mildly acidic environment. Heat labile enterotoxin may be released during Salmonella spp. growth, resulting in the loss of intestinal fluids. This enterotoxin is closely related functionally, immunologically and genetically to cholera toxin and the heat labile toxin of pathogenic E. coli (Jay et al., 2003). Most Salmonella strains also produce heat labile cytotoxin which may cause damage of the intestinal mucosal surface and general enteric symptoms and inflammation. For non-typhoidal Salmonella spp., infection is generally limited to a localised intestinal event. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 84 Dose response Human feeding trials for a range of Salmonella serovars were undertaken during the 1950’s to determine the relationship between the dose of pathogen ingested and the response of the individual (McCullough and Eisele.C.W, 1951a; McCullough and Eisele.C.W, 1951b; McCullough and Eisele.C.W, 1951c; McCullough and Eisele.C.W, 1951d). The study population consisted of healthy males confined in an institutional setting who were fed known doses of an individual Salmonella serovar. Infection was confirmed by recovering the administered Salmonella serovar from faecal samples. Fazil (1996) combined all the data from the feeding trials and found that a single Beta-Poisson relationship could adequately describe the dose-response for all serovars. However, a number of limitations exist on the use of such feeding trial data. Firstly the use of healthy adult male volunteers could underestimate the pathogenicity to the overall population. In addition, volunteers were exposed to high doses of Salmonella spp., with the minimum dose being 104 cells. In dose-response analysis, the critical region is the lower-dose region, as these are the doses that are most likely to exist in real food contamination events. This requires extrapolation of the model to doses much lower than those used in the human feeding trials. It must also be noted that the dose-response models are based on the risk of infection as an endpoint rather than illness, and therefore may introduce a level of conservatism into the dose-response relationship. It has been shown through salmonellosis outbreak investigations, that doses resulting in illnesses (gastroenteritis) were often several orders of magnitude lower than the doses reported in the feeding trials (D'Aoust, 1994). Using a reasonably large data set, the FAO/WHO in 2002 developed a dose-response model based on actual outbreak data. Although not subject to some of the inherent flaws associated with using purely experimental data, the data used in this model have a certain degree of uncertainty, which required assumptions to be made. This uncertainty is primarily due to the uncontrolled settings under which the information and data were collected. It is often difficult to determine the actual dose ingested (based on the level of the organism in the food at the time of consumption and the amount of food consumed), as well as determining the actual number of people exposed or ill during the outbreak. Figure 1: Uncertainty bounds for dose-response curves compared with expected value for the outbreak data (FAO/WHO, 2002). MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 85 Host factors Individual susceptibility to Salmonella spp. infection and/or disease can vary significantly, depending on host factors such as pre-existing immunity, nutrition, age, ability to elicit an immune response, structural and functional anomalies of the intestinal tract, or pre-existing disease (Gerba et al., 1996; Jay et al., 2003). Individuals who are generally at greater risk of infection and/or risk of developing more severe outcomes from exposure to Salmonella spp. include the very young, the elderly, pregnant women and the immunucompromised (organ transplant patients, cancer patients and AIDS patients) (Gerba et al., 1996). References Adak, G.K., Long, S.M. and O'Brien, S.J. (2002) Trends in indigenous foodborne disease and deaths, England and Wales: 1992 to 2000. Gut 51(6):832-841. Anon. (2003) Foodborne disease in Australia: incidence, notifications and outbreaks. Annual report of the OzFoodNet network, 2002. Commun Dis Intell. 27(2):209-243. Ashbolt, R., Givney, R., Gregory, J.E., Hall, G., Hundy, R., Kirk, M., McKay, I., Meuleners, L., Millard, G., Raupach, J., Roche, P., Prasopa-Plaizier, N., Sama, M.K., Stafford, R., Tomaska, N., Unicomb, L. and Williams, C. (2002) Enhancing foodborne disease surveillance across Australia in 2001: the OzFoodNet Working Group. Commun Dis Intell. 26(3):375-406. Bell, C. and Kyriakides, A. (2002) Salmonella: a practical approach to the organism and its control in foods. Blackwell Science, Oxford, UK. Blumer, C., Roche, P., Spencer, J., Lin, M., Milton, A., Bunn, C., Gidding, H., Kaldor, J., Kirk, M., Hall, R., Della-Porta, T., Leader, R. and Wright, P. (2003) Australia's notifiable diseases status, 2001: annual report of the National Notifiable Diseases Surveillance System. Commun Dis Intell. 27(1):1-78. Brenner, F.W., Villar, R.G., Angulo, F.J., Tauxe, R. and Swaminathan, B. (2000) Salmonella nomenclature. J Clin Microbiol 38(7):2465-2467. Bryan, F.L. and Doyle, M.P. (1995) Health risks and consequences of Salmonella and Campylobacter jejuni in raw poultry. J Food Prot. 58(3):326-344. Crerar, S.K., Nicholls, T.J. and Barton, M.D. (1999) Multi-resistant Salmonella Typhimurium DT104-implications for animal industries and the veterinary profession. Aust Vet J 77(3):170-171. D'Aoust, J.Y. (1994) Salmonella and the international food trade. Int J Food Microbiol 24(1-2):11-31. D'Aoust, J.Y. (1997) Salmonella species. In: Doyle, M.P., Beuchat, L.R., and Montville, T.J. eds. Food Microbiology: fundamentals and frontiers. ASM Press, Washington DC, USA, pp129-158. Dalton, C.B., Gregory, J., Kirk, M.D., Stafford, R.J., Givney, R., Kraa, E. and Gould, D. (2004) Foodborne disease outbreaks in Australia, 1995 to 2000. Commun Dis Intell. 28(2):211-224. FAO/WHO (2002) Risk Assessments of Salmonella in eggs and broiler chickens. Food and Agriculture Organization/World Health Organization, Geneva, Switzerland. ftp://ftp.fao.org/docrep/fao/005/y4392e/y4392e00.pdf. Fazil, A.M. (1996) A quantitative risk assessment model for salmonella. Drexel University, Philadelphia P.A, Accessed on 1996. Gerba, C.P., Rose, J.B. and Haas, C.N. (1996) Sensitive populations: who is at the greatest risk? Int J Food Microbiol 30(1-2):113-123. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 86 Hall, G.V. (2003) Gastroenteritis survey and foodborne disease in Australia 2002. Working paper series. National Centre for Epidemiology and Population Health, Australian National University, Canberra, ACT. Helms, M., Vastrup, P., Gerner-Smidt, P. and Molbak, K. (2003) Short and long term mortality associated with foodborne bacterial gastrointestinal infections: registry based study. BMJ 326(7385):357. Hohmann, E.L. (2001) Nontyphoidal salmonellosis. Clin Infect Dis 32(2):263-269. ICMSF. (1996) Salmonellae. In: Roberts, T.A., Baird-Parker, A.C., and Tompkin, R.B. eds. Microorganisms in foods 5: Microbiological specifications of food pathogens. First ed, Blackie Academic & Professional, London, UK, pp217-264. Jay, S., Davos, D., Dundas, M., Frankish, E. and Lightfoot, D. (2003) Salmonella. In: Hocking, A.D. eds. Foodborne microorganisms of public health significance. 6 ed, Australian Institute of Food Science and Technology, NSW Branch, Food Microbiology Group, Waterloo, NSW, pp207-266. McCullough, N. and Eisele.C.W. (1951a) Experimental human salmonellosis. I. Pathogenicity of strains of Salmonella meleagridis and Salmonella anatum obtained from spray-dried whole egg. J Infect Dis 88(3):278289. McCullough, N. and Eisele.C.W. (1951b) Experimental human salmonellosis. II. Immunity studies following experimental illness with Salmonella meleagridis and Salmonella anatum. J Immunol. 66(5):595-608. McCullough, N. and Eisele.C.W. (1951c) Experimental human salmonellosis. III. Pathogenicity of strains of Salmonella newport, Salmonella derby, and Salmonella bareilly obtained from spray-dried whole egg. J Infect Dis 89(3):209-213. McCullough, N. and Eisele.C.W. (1951d) Experimental human salmonellosis. IV. Pathogenicity of strains of Salmonella pullorum obtained from spray-dried whole egg. J Infect Dis 89(3):259-265. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M. and Tauxe, R.V. (1999) Food-related illness and death in the United States. Emerg.Infect Dis 5(5):607-625. Roche, P., Spencer, J., Lin, M., Gidding, H., Kirk, M., Eyeson-Annan, M., Milton, A., Witteveen, D. and Merianos, A. (2001) Australia's notifiable diseases status, 1999: annual report of the National Notifiable Diseases Surveillance System. Commun Dis Intell. 25(4):190-245. Yohannes, K., Roche, P., Blumer, C., Spencer, J., Milton, A., Bunn, C., Gidding, H., Kirk, M. and Della-Porta, T. (2004) Australia's notifiable diseases status, 2002: Annual report of the National Notifiable Diseases Surveillance System. Commun Dis Intell. 28(1):6-68. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 87 ANNEX 5: Bulk milk tank prevalence This annex presents a summary of the literature sources used to establish the prevalence of pathogenic bacteria in on-farm bulk milk tanks. Exclusion criteria included geographical considerations, the age profile of the herds, sample size and analytical methodology. The geographical location of the studies that were accepted was limited to those countries that were considered to have production systems similar to those found in Australia. Five geographical regions were considered: Australia and New Zealand, North America, UK/Ireland, continental Europe and Scandinavia. Within these, countries consideration was also made about the production systems. For example, results from Scandinavian countries where cows were housed during the winter months were generally excluded. For the four pathogens considered, Campylobacter spp., enterohaemorrhagic Escherichia coli, Salmonella spp. and Listeria monocytogenes, there is a separate summary table. This includes information on the source of the data, the country, the total number of samples, the number of samples where pathogens were detected and the prevalence. In addition, there is a plot of the cumulative probability distribution of the herd prevalence based on the data presented in each table. The method used for developing a combined cumulative distribution for the herd prevalence was by using weighted average of the cumulative percentiles (Vose, 2008). In this case all studies are given equal weighting. The method is briefly outlined below: • For each study the number of detections (positives) and the total number of samples are found • The parameters of a beta distribution, Beta(α1, α2) with parameters, α1 = s + 1 and α2 = n – s + 1 (assuming a uniform prior) are calculated • In the statistical program, R (www.r-project.org) the cumulative percentiles for the corresponding Beta distribution is determined using the pbeta command • The cumulative distributions for each Beta distribution are then average to give an overall cumulative distribution • The resulting distribution is entered into @Risk as a cumulative distribution with a lower limit of zero, and an upper limit judged on the cumulative distribution values MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 88 Campylobacter spp. Author/Publication Reference Country SA unpublished (1995-2002) 1 Australia WA unpublished (2007) 2 Australia Bachmann and Spahr (1995) 3 Switzerland Davidson et al. (1989) 4 Canada Desmasures et al. (1997) 5 France Doyle and Roman (1982) 6 USA Jayarao and Henning (2001)* 7 USA Jayarao et al. (2006) 8 USA Lovett et al.(1983) 9 USA McManus and Lanier (1987) 10 USA Oosterom et al. (1982) 11 Netherlands Rohrbach et al. (1992)* 12 USA Steele et al. (1997) 13 Canada Food Standards Agency (2002) 14 UK Food Standards Agency (2003) 15 UK Whyte et al. (2004) 16 Ireland Wyatt and Timm (1982) 17 USA * Statistically different to the WA unpublished (2007) prevalence Samples (n) 95 183 496 192 69 108 131 248 195 237 200 292 1720 602 610 62 50 Positive (s) 0 0 0 3 1 1 12 5 3 1 0 36 8 5 5 1 0 Prevalence 0 0 0 0.016 0.014 0.009 0.092 0.02 0.015 0.004 0 0.123 0.005 0.008 0.008 0.016 0 Samples 118 500 143 1802 603 131 248 859 100 115 1720 602 610 Positive 0 0 1 171 0 5 6 36 0 11 15 1 1 Prevalence 0 0 0.007 0.095 0 0.038 0.024 0.042 0 0.096 0.009 0.002 0.002 E. coli (EHEC) Author/Publication Reference Country WA unpublished (2007) 1 Australia Coia et al. (2001) 2 Scotland de Rue et al. (2004) 3 Belgium Desmarchelier (1997) 4 Australia Hancock et al. (1994) 5 USA Jayarao and Henning (2001) 6 USA Jayarao et al. (2006) 7 USA Karns et al. (2007) 8 USA Massa et al. (1999) 9 Italy Padhye and Doyle (1991)* 10 USA Steele et al. (1997) 11 Canada Food Standards Agency (2002) 12 UK Food Standards Agency (2003) 13 UK * Statistically different to the WA unpublished (2007) prevalence Salmonella spp. Author/Publication SA unpublished (1995-2002) WA unpublished (2007) Bachmann and Spahr (1995)* de Rue et al. (2004) Desmasures et al. (1997) Eliskases-Lechner and Ginzinger (1999) Humphrey and Hart (1988) Humphrey and Hart (1988)* Jayarao et al. (2006) Murinda et al. (2002a) O'Donnell (1995)* Reference 1 2 3 4 5 Country Australia Australia Switzerland Belgium France Samples 108 183 456 143 69 Positive 4 14 0 0 2 Prevalence 0.037 0.077 0 0 0.029 6 Austria 201 0 0 153 985 248 268 0 2 15 6 0 0.002 0.06 0.022 1673 6 0.004 292 1720 861 602 610 26 3 22 2 2 0.089 0.002 0.026 0.003 0.003 7 8 9 10 UK UK USA USA England and 11 Wales Rohrbach et al. (1992) 12 USA Steele et al. (1997)* 13 Canada van Kessel et al. (2004) 14 USA Food Standards Agency (2002)* 15 UK Food Standards Agency (2000)* 16 UK * Statistically different to the WA unpublished (2007) prevalence MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 89 Listeria monocytogenes Author/Publication SA unpublished (1995-2002) WA unpublished (2007) Bachmann and Sphar (1995) Davidson et al. (1989) de Rue et al. (2004) Desmasures et al. (1997) Doores and Amelang (1988) Eliskases-Lechner and Ginzinger (1989) Fenlon et al. (1995) Husu (1990) Ibrahim and MacRae (1991) Jayarao and Henning (2001) Jayarao et al. (2006) Kongo et al. (2006) Liewen and Platz (1988) Lovett et al.(1987) Meyer-Brosetta et al. (2003) Muraoka et al. (2003) Navratilova et al. (2004) O'Donnell (1995) Reference 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Country Australia Australia Switzerland Canada Belgium France USA Austria Scotland Finland Australia USA USA Portugal USA USA France USA Czech Republic England and Wales USA Canada USA UK UK Sweden Samples 97 183 340 192 143 69 2511 201 638 314 150 131 248 105 200 650 1459 474 Positive 0 0 2 2 9 4 71 0 43 7 0 6 3 2 8 27 35 23 Prevalence 0 0 0.006 0.01 0.063 0.058 0.028 0 0.067 0.022 0 0.046 0.012 0.019 0.04 0.042 0.024 0.049 278 6 0.022 2009 102 0.051 292 1720 861 602 610 294 12 47 56 101 103 3 0.041 0.027 0.065 0.168 0.169 0.01 Rohrbach et al. (1992) 21 Steele et al. (1997) 22 van Kessel et al. (2004) 23 Food Standards Agencu (2002)* 24 Food Standards Agency (2003)* 25 Waak et al. (2002) 26 * Statistically significant different to the WA unpublished (2007) prevalence MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 90 C a m p yl o b ac ter spp. 0.20 0.05 0.00 10 15 2 4 6 8 10 Reference S al m on el l a spp. Listeria monocytogenes 12 0.10 0.15 0.20 ** * * * * * 5 10 15 0.00 0.05 0.10 Prevalence 0.15 0.20 Reference 0.05 0.00 0.15 Prevalence 0.15 0.10 0.00 5 Prevalence * 0.10 * * 0.05 Prevalence 0.20 EHEC 0 5 Reference 10 15 20 25 Reference Prevalence of Campylobacter spp., EHEC, Salmonella spp. and L. monocytogenes in bulk milk tanks. Error bars represent the 95% confidence intervals. Results highlighted with an asterisk (*) indicate that the prevalence is significantly different from the WA unpublished (2007) raw milk data. Details of each numbered reference can be found in the corresponding tables in this annex. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 91 Annex 6: With-in herd prevalence This annex presents a summary of the literature sources used for with-in herd prevalence that were accepted after a preliminary review. With-in herd prevalence was largely based on the detection of pathogens in faecal samples. Exclusion criteria included geographical considerations, the age profile of the herds, sample size and analytical methodology. The geographical location of the studies that were accepted was limited to those countries that were considered to have production systems similar to those found in Australia. Five geographical regions were considered: Australia and New Zealand, North America, UK/Ireland, continental Europe and Scandinavia. Within these, countries consideration was also made about the production systems. For example, results from Scandinavian countries where cows were housed during the winter months were generally excluded. For the four pathogens considered, Campylobacter spp., enterohaemorrhagic Escherichia coli, , Salmonella spp. and Listeria monocytogenes, there is a separate summary table. This includes information on the source of the data, the country, the total number of samples, the number of samples where pathogens were detected and the prevalence. In addition, there is a plot of the cumulative probability distribution of the herd prevalence based on the data presented in each table. The method used for developing a combined cumulative distribution for the herd prevalence was by using weighted average of the cumulative percentiles (Vose, 2008). In this case all studies are given equal weighting. The method is briefly outlined below: • For each study the number of detections (positives) and the total number of samples are found • The parameters of a beta distribution, Beta(α1, α2) with parameters, α1 = s + 1 and α2 = n – s + 1 (assuming a uniform prior) are calculated • In the statistical program, R (www.r-project.org) the cumulative percentiles for the corresponding Beta distribution is determined using the pbeta command • The cumulative distributions for each Beta distribution are then average to give an overall cumulative distribution • The resulting distribution is entered into @Risk as an cumulative distribution with a lower limit of zero, and an upper limit judged on the cumulative distribution values Campylobacter spp. Author (Bae et al., 2005) (Bailey et al., 2003) (Beumer et al., 1988) (Dodson and LeJeune, 2005) (Doyle and Roman, 1982) (Hakkinen et al., 2007) (Harvey et al., 2004) (Murinda et al., 2004) (Oosterom et al., 1982) (Sato et al., 2004) (Wesley et al., 2000) Reference 1 2 3 4 5 6 7 8 9 10 11 Country USA Australia Netherlands USA USA Finland USA USA Netherlands USA USA MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK Samples 311 150 904 686 78 952 720 411 200 1191 2085 Positive 97 64 41 48 50 186 20 5 11 234 786 Prevalence 0.312 0.427 0.045 0.07 0.641 0.195 0.028 0.012 0.055 0.196 0.377 92 E. coli (EHEC) Author (Cobbold and Desmarchelier, 2000) (Doane et al., 2007) (Dodson and LeJeune, 2005) (Edrington et al., 2004) (Fegan et al., 2004c) (Garber et al., 1999) (Hallaran and Sumner, 2001) (Hancock et al., 1994) (Lejeune et al., 2006) (Lejeune et al., 2006) (Mechie et al., 1997) (Murinda et al., 2002b) (Pradel et al., 2000) (Renter et al., 2005) (Rice et al., 1997) (Rugbjerg et al., 2003) (Wells et al., 1991) (Zschock et al., 2000) Reference 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Country Australia USA USA USA Australia USA Australia USA USA Norway UK USA France USA USA Denmark USA Germany Samples 120 408 1026 720 310 4361 505 3570 750 680 2786 415 471 246 397 560 662 726 Positive 2 16 21 52 39 52 1 10 5 0 63 8 162 33 18 14 1 131 Prevalence 0.017 0.039 0.02 0.072 0.126 0.012 0.002 0.003 0.007 0 0.023 0.019 0.344 0.134 0.045 0.025 0.002 0.18 Reference 1 2 3 4 5 6 7 Country USA USA Australia Australia USA USA USA Samples 580 720 310 606 91 6861 415 Positive 39 262 21 157 25 145 9 Prevalence 0.067 0.364 0.068 0.259 0.275 0.021 0.022 Reference 1 2 3 Country Australia Canada Finland Samples 150 401 3878 Positive 4 21 373 Prevalence 0.027 0.052 0.096 Salmonella spp. Author (Dodson and LeJeune, 2005) (Edrington et al., 2004) (Fegan et al., 2004b) (Fegan et al., 2005b) (Kabagambe et al., 2000) (Losinger et al., 1995) (Murinda et al., 2002a) L. monocytogenes Author (Bailey et al., 2003) (Fedio and Jackson, 1992) (Husu, 1990) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 93 C a m p yl o b ac ter spp. 0.8 0.6 0.2 0.0 4 6 8 10 5 10 15 Reference S al m on el l a spp. Listeria monocytogenes 0.6 0.4 0.2 0.0 0.0 0.2 0.4 Prevalence 0.6 0.8 Reference 0.8 2 Prevalence 0.4 Prevalence 0.4 0.0 0.2 Prevalence 0.6 0.8 EHEC 1 2 3 4 5 6 7 1 Reference 2 3 Reference With-in herd prevalence of Campylobacter spp., EHEC, Salmonella spp. and L. monocytogenes in dairy cows. Error bars represent the 95% confidence intervals. Results highlighted with an asterisk (*) indicate that the prevalence is significantly different from the WA unpublished (2007) raw milk data. Details of each numbered reference can be found in the corresponding table in this appendix. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 94 Annex 7: Growth and inactivation rates equations Enterohaemorragic Escherichia coli (EHEC) The growth rate of EHEC is considered to be identical as that of non-pathogenic E. coli based on the experimental results from Salter (1998). Ross et al. (2003) proposed an extension of the Square-root model to include factors such as temperature, water activity( aw), pH and lactic acid concentration [LAC] on the relative growth rate, µ (Equation 1): Equation 1 Table 1: Growth model equation parameters for E. coli (Ross et al., 2003) Parameter c Tmin Tmax pHmin pHmax Value 0.2345 4.14 49.55 3.909 8.86 Parameter Umin Dmin awmin d pKa Value 10.43 995.508 0.9508 0.2636 3.86 Campylobacter spp. Campylobacter spp. including C. jejuni are microaerotolerant and are not capable of growth in raw milk. Doyle and Roman (1982) performed experimental studies into the survival of eight different strains of C. jejuni in unpasteurised milk stored at 4 °C. Initial inoculum levels of >107cfu/ml were used. Results of this study showed that there are considerable differences in survival between strains of C. Jejuni. Some strains were not detected after 9 days (6 log decline), while the concentrations of other strains declined by only 1 log in the same time. A graphical summary of the experimental trials is presented below: MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 95 5 10 15 20 5 10 15 20 6 4 0 2 log 10 C. jejuni 8 6 4 0 0 0 5 10 15 20 0 5 10 15 20 Time (days ) Time (day s ) Tria l 5 Tria l 6 Trial 7 Trial 8 10 15 20 Time (day s) 0 5 10 15 Time (day s) 20 6 4 0 2 log10 C. jejuni 8 6 4 0 2 log10 C. jejuni 6 4 0 2 log10 C. jejuni 8 6 4 2 5 8 Time (day s) 8 Time (day s) 0 0 Trial 4 2 log 10 C. jejuni 6 4 0 2 log 10 C. jejuni 8 6 4 2 0 log 10 C. jejuni 0 log10 C. jejuni Trial 3 8 Tria l 2 8 Tria l 1 0 5 10 15 20 Time (days ) 0 5 10 15 20 Time (day s ) In order to capture the between-strain variability a non-linear mixed effects model was fitted to the experimental data: log10N ~ (β0 + ai) – exp(β1 + bi) time where β0 is the common intercept, β1 is a common slope and ai and bi are the random effects for the intercept and slope, respectively. The slope parameter, β1 is the logarithm of the firstorder inactivation rate. Parameter β0 β1 Standard deviation of ai Standard deviation of bi Value (Standard error) 6.898379 (0.2101275) -0.878915 (0.1985563) 3.575302x10-5 0.5040929 The first-order rate constant for the inactivation of Campylobacter spp. in raw milk was coded into @Risk as: EXP(RiskNormal(-0.8789,0.5041)), using the expected value of β1 and the standard deviation of the random effect bi. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 96 Salmonella spp. The growth rate model for Salmonella spp. is based on the mixed cocktail study by Gibson et al. (1988) in broth culture. The original growth curve data was obtained from the Combase predictive microbiology database (www.combase.cc). A modified form the Baranyi model: was fitted to the experimental growth curves using the statistical program R Version 2.6.2 (R Development Core Team, 2007) to determine the maximum specific growth rate, μm. A quadratic surface model including temperature, pH and added salt concentration was fitted to the specific growth rate data: Adjusted R2 = 0.97, Residual standard error = 0.1915 on 58 degrees of freedom Listeria monocytogenes The growth rate for L. monocytogenes was derived from data presented in Xanthiakos et al. (2006). As there is only sub-optimal temperature data the Square root model was used to predict the effect of temperature of growth rates: 0.35 0.3 Sqrt(growth rate) 0.25 0.2 0.15 0.1 0.05 0 0 2 4 6 8 10 12 14 16 18 Temperature (°C) MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 97 ANNEX 8: Risk management questions 1. What are the risks to public health and safety posed by the consumption, in Australia, of raw cow milk? 2. What are the factors that would have the greatest impact on public health and safety along the production chain for raw cow milk for direct consumption? Specific questions posed by the project team in relation to raw cow milk for human consumption are: 1) What are the microbial hazards of public health significance in raw milk? Are there hazards specific to particular species, i.e. cow milk? 2) What are the prevalence and levels of identified hazards in raw cow milk? 3) Do these levels pose a risk if the raw cow milk is used for direct consumption? 4) What are the factors during primary production that impact on the level of these hazards? 5) What practices/controls have the greatest impact on the level of hazard? 6) What is the impact of retail and consumer handling on the level of risk to public health and safety on these hazards? MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 98 Annex 9: References Adesiyun, A.A. (1994) Bacteriological quality and associated public health risk of preprocessed bovine milk in Trinidad. Int J Food Microbiol 21(3):253-261. Adesiyun, A.A., Webb, L. and Rahaman, S. (1995) Microbiological quality of raw cow's milk at collection centers in Trinidad. Journal of Food Protection 58(2):139-146. Adesiyun, A.A., Webb, L.A., Romain, H. and Kaminjolo, J.S. (1997) Prevalence and characteristics of strains of Eschericia coli isolated from milk and feces of cows on dairy farms in Trinidad. Journal of Food Protection 60(10):1174-1181. Allerberger, F. and Guggenbichler, J.P. (1989) Listeriosis in Austria--report of an outbreak in 1986. Acta Microbiol Hung. 36(2-3):149-152. Allerberger, F., Wagner, M., Schweiger, P., Rammer, H.P., Resch, A., Dierich, M.P., Friedrich, A.W. and Karch, H. (2001) Escherichia coli O157 infections and unpasteurised milk. Euro.Surveill 6(10):147-151. Anon. (2003) Foodborne Microorganisms of Public Health Significance. In: Hocking, A.D. eds. 6th edition ed, Australian Institute of Food Science and Technology Incorporated. Anon (1984) Kindergarten field trip to a farm, June 25, 1984, Vancouver Island. Disease Surveillance, British Colunbia 1984(5):201-203. Ansay, S.E. and Kaspar, C.W. (1997) Survey of retail cheeses, dairy processing environments and raw milk for Escherichia coli O157:H7. Lett.Appl.Microbiol. 25(2):131-134. Aygun, O. and Pehlivanlar, S. (2006) Listeria spp in the raw milk and dairy products in Antakya, Turkey. Food Control 17:676-679. Bachmann, H.P. and Spahr, U. (1995) The fate of potentially pathogenic bacteria in Swiss hard and semihard cheeses made from raw milk. J Dairy Sci 78(3):476-483. Bae, W., Kaya, K.N., Hancock, D.D., Call, D.R., Park, Y.H. and Besser, T.E. (2005) Prevalence and anitmicrobial resistance of thermophilic Campylobacter spp. from cattle farms in Washington state. Applied and Environmental Microbiology 71(1):169-174. Bailey, G.D., Vanselow, B.A., Hornitzky, M.A., Hum, S.I., Eamens, G.J., Gill, P.A., Walker, K.H. and Cronin, J.P. (2003) A study of the foodborne pathogens: Campylobacter, Listeria and Yersinia, in faeces from slaughter-age cattle and sheep in Australia. Commun Dis Intell. 27(2):249-257. Beumer, R.R., Cruysen, J.J. and Birtantie, I.R. (1988) The occurrence of Campylobacter jejuni in raw cows' milk. J Appl Bacteriol. 65(2):93-96. Black, R.E., Levine, M.M., Clements, M.L., Hughes, T.P. and Blaser, M.J. (1988) Experimental Campylobacter jejuni infection in humans. J Infect Dis 157(3):472-479. Blaser, M.J., Cravens, J., Powers, B.W., Laforce, F.M. and Wang, W.L. (1979) Campylobacter enteritis associated with unpasteurized milk. Am J Med 67(4):715-718. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 99 Carlos, V.S., Oscar, R.S. and Irma, Q.R.E. (2001) Occurrence of Listeria species in raw milk in farms on the outskirts of Mexico city. Food Microbiology 18(2):177-181. CDC (2002) Foodborne Outbreak Response Surveillance Unit. http://www2a.cdc.gov/ncidod/foodborne/fbsearch.asp. Accessed on 10 May 2005. Chen, Y., Ross, W.H., Gray, M.J., Wiedmann, M., Whiting, R.C. and Scott, V.N. (2006) Attributing risk to Listeria monocytogenes subgroups: Dose response in relation to genetic lineages. Journal of Food Protection 69(2):335-344. Cho, S., Bender, J.B., Diez-Gonzalez, F., Fossler, C.P., Hedberg, C.W., Kaneene, J.B., Ruegg, P.L., Warnick, L.D. and Wells, S.J. (2006) Prevalence and characterisation of Escherichia coli O157 isolates from Minnesota dairy farms and county fairs. Journal of Food Protection 69(2):252-259. Chye, F.Y., Abdullah, A. and Ayob, M.K. (2004) Bacteriological quality and safety of raw milk in Malaysia. Food Microbiology 21(5):535-541. Clough, H.E., Damian, C. and French, N.P. (2009) Quantifying exposure to Verocytotoxigenic Escherichia coli O157 in milk sold as pasteurized: A model-based approach. International Journal of Food Microbiology 131:95-105. Cobbold, R. and Desmarchelier, P. (2000) A longitudinal study of Shiga-toxigenic Escherichia coli (STEC) prevalence in three Australian diary herds. Vet Microbiol 71(12):125-137. Cobbold, R. and Desmarchelier, P. (2001) Characterisation and clonal relationships of Shigatoxigenic Escherichia coli (STEC) isolated from Australian dairy cattle. Vet Microbiol 79(4):323-335. Cobbold, R.N., Rice, D.H., Szymanski, M., Call, D.R. and Hancock, D.D. (2004) Comparison of shiga-toxigenic Escherichia coli prevalences among dairy, feedlot, and cow-calf herds in Washington State. Appl Environ Microbiol 70(7):4375-4378. Coia, J.E., Johnston, Y., Steers, N.J. and Hanson, M.F. (2001) A survey of the prevalence of Escherichia coli O157 in raw meats, raw cow's milk and raw-milk cheeses in south-east Scotland. International Journal of Food Microbiology.2001 66(1/2):63-69. Conedera, G., Dalvit, P., Martini, M., Galiero, G., Gramaglia, M., Goffredo, E., Loffredo, G., Morabito, S., Ottaviani, D., Paterlini, F., Pezzotti, G., Pisanu, M., Semprini, P. and Caprioli, A. (2004) Verocytotoxin-producing Escherichia coli O157 in minced beef and dairy products in Italy. International Journal of Food Microbiology.2004 96(1):67-73. Cook, G.M. and Sandeman, R.M. (2000) Sources and characterisation of spore-forming bacteria in raw milk. Australian Journal of Dairy Technology 55(3):119-126. Cools, I., D’Haese, E., Uyttendaele, M., Storms, E., Nelis, H.J., Debevere, J. (2005) Solid phase cytometry as a tool to detect viable but non-culturable cells of Campylobacter jejuni. Journal of Microbiological Methods 63(2):107-114. Dairy Australia (2004) Australian Dairy Industry in Focus 2004. Dairy Australia. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 100 Dairy Australia (2006) Personal Communication. Davidson, R.J., Sprung, D.W., Park, C.E. and Rayman, M.K. (1989) Occurrence of Listeria monocytogenes, Campylobacter spp and Yersinia enterocolitica in Manitoba raw milk. Canadian Institute of Food Science and Technology Journal 22(1):70-74. De Louvois, J. and Rampling, A. (1998) One fifth of samples of unpasteurised milk are contaminated with bacteria. British Medical Journal 316:625. De Rue, K., Grijspeerdt, K. and Herman, L. (2004) A Belgian survey of hygiene indicator bacteria and pathogenic bacteria in raw milk and direct marketing of raw milk farm products. Journal of Food Safety 24(1):17-36. Desmarchelier, P.M. (1997) Enterohemorrhagic Escherichia coli - the Australian perspective. Journal of Food Protection 60(11):1447-1450. Desmarchelier, P.M. and Fegan, N. (2003) Enteropathogenic Escherichia coli. In: Hocking, A.D. eds. Foodborne microorganisms of public health significance. Sixth ed, Australian Institute of Food Science and Technology (NSW Branch), Sydney, Australia, pp. 267-310. Desmasures, N. (1997) Microbiological composition of raw milk from selected farms in the Camembert region of Normandy. Journal-of-Applied-Microbiology 83(1):53-58. Desmasures, N., Bazin, F. and Gueguen, M. (1997) Microbiological composition of raw milk from selected farms in the Camembert region of Normandy. J Appl Microbiol 83(1):53-58. Doane, C.A., Pangloli, P., Richards, H.A., Mount, J.R., Golden, D.A. and Draughon, F.A. (2007) Occurrence of Escherichia coli O157:H7 in diverse farm environments. Journal of Food Protection 70(1):6-10. Dodson, K. and LeJeune, J. (2005) Escherichia coli O157 : H7, Campylobacter jejuni, and Salmonella prevalence in cull dairy cows marketed in northeastern Ohio. Journal of Food Protection 68(5):927-931. Doores, S. and Amelang, J. (1988) Incidence of Listeria monocytogenes in the raw milk supply in Pennsylvania. In: IFT Annunal Meeting, IFT Ann. Meeting, 19-22 June., FT Chicago, New Orleans, pp157. Doyle, M.P. and Roman, D.J. (1982) Prevalence and survival of Campylobacter jejuni in unpasteurized milk. Appl Environ Microbiol 44(5):1154-1158. Doyle, M.P., Zhao, T., Meng, J. and Zhao, S. (1997) Escherichia coli O157H:7. In: Doyle, M.P., Beuchat, L.R., and Montville, T.J. eds. Food Microbiology: Fundamentals and frontiers. ASM Press, Washington, USA, pp. 171-191. Edrington, T.S., Hume, M.E., Looper, M.L., Schultz, C.L., Fitzgerald, A.C., Callaway, T.R., Genovese, K.J., Bischoff, K.M., McReynolds, J.L., Anderson, R.C. and Nisbet, D.J. (2004) Variation in the faecal shedding of Salmonella and E.coli O157:H7 in lactating dairy cattle and examination of Salmonella genotypes using pulsed-field gel elctrophoresis. Letters in Applied Microbiology 38:366-372. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 101 Eliskases-Lechner, F. and Ginzinger, W. (1999) Occurrence of Listeria and Salmonella in raw milk from farms without feeding silage. Ernaehrung.1999 23(9):356-358. European Commission (2003) Trends and Sources of Zoonotic Agents in animals, feedingstuffs, food and man in the European Union and Norway in 2001. In: Community Reference Laboratory on the Epidemiology of Zoonoses. eds. Report No. SANC/56/2003 Part 1, CRL, Berlin, Germany. http://ec.europa.eu/food/food/biosafety/salmonella/zoonoses_reps_2001_en.htm. European Commission (2004) Trends and Sources of Zoonotic Agents in animals, feedingstuffs, food and man in the European Union and Norway in 2002. In: Community Reference Laboratory on the Epidemiology of Zoonoses. eds. Report No. SANC/29/2004 Part 1, CRL, Berlin, Germany. http://ec.europa.eu/food/food/biosafety/salmonella/zoonoses_reps_2002_en.print.htm. Evans, M.R., Roberts, R.J., Ribeiro, C.D., Gardner, D. and Kembrey, D. (1996) A milk-borne campylobacter outbreak following an educational farm visit. Epidemiol Infect 117(3):457462. FAO/WHO (2006) Benefits and potential risks of the lactoperoxidase system of raw milk preservation. FAO/WHO (2004) Risk assessment of Listeria monocytogenes in ready-to-eat foods. Technical report. Farber, J.M., Sanders, G.W. and Malcolm, S.A. (1988) The presence of Listeria spp. in raw milk in Ontario. Canadian Journal of Microbiology 34:95-100. Fedio, W.M. and Jackson, H. (1990) Incidence of Listeria monocytogenes in raw bulk milk in Alberta. Canadian Institute of Food Science and Technology Journal 23(4/5):236-238. Fedio, W.M. and Jackson, H. (1992) On the origin of Listeria monocytogenes in raw bulktank milk. International Dairy Journal.1992 2(3):197-208. Fedio, W.M., Schoonderwoerd, M., Shute, R.H. and Jackson, H. (1990) A case of bovine mastitis caused by Listeria monocytogenes. Canadian Veterinary Journal 31:773-775. Fegan, N., Higgs, G., Vanderlinde, P. and Desmarchelier, P. (2004a) Enumeration of Escherichia coli O157 in cattle faeces using most probable number technique and automated immunomagnetic separation. Letters in Applied Microbiology 38(1):56-59. Fegan, N., Vanderlinde, P., Higgs, G. and Desmarchelier, P. (2004b) Quantification and prevalence of Salmonella in beef cattle presenting at slaughter. Journal of Applied Microbiology 97(5):892-898. Fegan, N., Higgs, G., Vanderlinde, P. and Desmarchelier, P. (2005a) An investigation of Escherichia coli O157 contamination of cattle during slaughter at an abattoir. Journal of Food Protection 68(3):451-457. Fegan, N., Vanderlinde, P., Higgs, G. and Desmarchelier, P. (2005b) A study of the prevalence and enumeration of Salmonella enterica in cattle and on carcasses during processing. Journal of Food Protection 68(6):1147-1153. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 102 Fegan, N., Vanderlinde, P., Higgs, G. and Desmarchelier, P. (2004c) The prevalence and concentration of Escherichia coli O157 in faeces of cattle from different production systems at slaughter. J Appl Microbiol 97(2):362-370. Fenlon, D.R., Stewart, T. and Donachie, W. (1995) The incidence, numbers and types of Listeria monocytogenes isolated from farm bulk tank milks. Lett Appl Microbiol 20(1):57-60. Fenlon, D.R. and Wilson, J. (1989) The incidence of Listeria monocytogenes in raw milk from farm bulk tanks in north-east Scotland. J Appl Bacteriol. 66(3):191-196. Food Safety Authority of Ireland (2004) Report on zoonoses in Ireland 2000 and 2001. Food Safety Authority of Ireland. Food Standards Agency (2003) Report on the national study of the microbiological quality and heat processing of cows' milk. www.food.gov.uk/multimedia/pdfs/milksurvey.pdf . Accessed on 11 March 2006. Food Standards Agency (2002) Zoonoses Report United Kingdom 2000. Department for Environment, Food and Rural Affairs, London. Francis, A.J., Nimmo, G.R., Efstratiou, A., Galanis, V. and Nuttall, N. (1993) Investigation of milk-borne Streptococcus zooepidemicus infection associated with glomerulonephritis in Australia. J Infect 27(3):317-323. FSANZ (2006) A Risk Profile of Dairy Products in Australia. Food Standards Australia New Zealand. http://www.foodstandards.gov.au/standardsdevelopment/proposals/proposalp296primaryp280 6.cfm. Garber, L., Wells, S., Schroeder-Tucker, L. and Ferris, K. (1999) Factors associated with fecal shedding of verotoxin-producing Escherichia coli O157 on dairy farms. J Food Prot. 62(4):307-312. Garber, L.P., Wells, S.J., Hancock, D.D., Doyle, M.P., Tuttle, J., Shere, J.A. and Zhao, T. (1995) Risk factors for fecal shedding of Escherichia coli O157:H7 in dairy calves. J Am Vet Med Assoc 207(1):46-49. Gaya, P., Sanchez, J., Medina, M. and Nunez, M. (1998) Incidence of Listeria monocytogenes and other Listeria species in raw milk produced in Spain. Food Microbiology.1998 15(5):551555. Gibson, A.M., Bratchell, N. and Roberts, T.A. (1988) Predicting microbial growth: growth responses of Salmonellae in a laboratory medium as affected by pH, sodium chloride and storage temperature. International Journal of Food Microbiology 6:155-178. Gran, H.M., Wetlesen, A., Mutukumira, A.N., Rukure, G. and Narvhus, J.A. (2003) Occurrence of pathogenic bacteria in raw milk, cultured pasteurised milk and naturally soured milk produced at small-scale dairies in Zimbabwe. Food Control 14(8):539-544. Grau, F.H., Brownlie, L.E. and Roberts, E.A. (1968) Effect of some preslaughter treatments on the Salmonella population in the bovine rumen and faeces. Journal Applied Bacteriology 31(1):157-163. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 103 Grau, F.H., Brownlie, L.E. and Smith, M.G. (1969) Effects of food intake on numbers of Salmonellae and Escherichia coli in rumen and faeces of sheep. Journal Applied Bacteriology 32(1):112-117. Hakkinen, M., Heiska, H. and Hanninen, M.L. (2007) Prevalence of Campylobacter spp in cattle in Finland and antimicrobial susceptibilities of bovine Campylobacter jejuni strains. Applied and Environmental Microbiology 73(10):3232-3238. Hallaran, G. and Sumner, J. (2001) Prevalence of E coli O157 in dairy cows presented for slaughter in Victoria. Australian Veterinary Journal 79(10):707-708. Hancock, D.D., Besser, T.E., Kinsel, M.L., Tarr, P.I., Rice, D.H. and Paros, M.G. (1994) The prevalence of Escherichia coli O157.H7 in dairy and beef cattle in Washington State. Epidemiol Infect 113(2):199-207. Harrington, P., Archer, J., Davis, JP., Croft, DR. and Varma, JK. (2001) Outbreak of Campylobacter jejuni Infections Associated with Drinking Unpasteurized Milk Procured through a Cow-Leasing Program --- Wisconsin, 2001. Morbidity and Mortality 51(25):548549. Harvey, J. and Gilmour, A. (1992) Occurrence of Listeria species in raw milk and dairy products produced in Northern Ireland. J Appl Bacteriol. 72(2):119-125. Harvey, R.B., Droleskey, R.E., Sheffield, C.L., Edrington, T.S., Callaway, T.R., Anderson, R.C., Drinnon, D.L.J., Ziprin, R.L., Scott, H.M. and Nisbet, D.J. (2004) Campylobacter prevalence in lactating dairy cows in the United States. Journal of Food Protection 67(7):1476-1479. Hassan, L., Mohammed, H.O., McDonough, P.L. and Gonzalez, R.N. (2000) A crosssectional study on the prevalence of Listeria monocytogenes and Salmonella in New York dairy herds. J Dairy Sci 83(11):2441-2447. Hayes, P.S., Feeley, J.C., Graves, L.M., Ajello, G.W. and Fleming, D.W. (1986) Isolation of Listeria monocytogenes from raw milk. Appl Environ Microbiol 51(2):438-440. Headrick, M.L., Korangy, S., Bean, N.H., Angulo, F.J., Altekruse, S.F., Potter, M.E. and Klontz, K.C. (1998) The epidemiology of raw milk-associated foodborne disease outbreaks reported in the United States, 1973 through 1992. Am J Public Health 88(8):1219-1221. Headrick, M.L., Timbo, B., Klontz, K.C. and Werner, S.B. (1997) Profile of raw milk consumers in California. Public Health Rep 112(5):418-422. Heuvelink, A.E., Bleumink, B., Biggelaar, F.L.A.M., Te-Giffel, M.C., Beumer, R.R. and Boer, E.d. (1998) Occurrence and survival of verocytotoxin-producing Escherichia coli O157 in raw cow's milk in the Netherlands. Journal of Food Protection.1998 61(12):1597-1601. Holko, I., Urbanova, J., Kantikova, M., Pastorova, K. and Kmet, V. (2002) PCR detection of Listeria monocytogenes in milk and milk products and differentiation of suspect isolates. Acta Veterinaria Brno 71(1):125-131. Humphrey, T.J. and Beckett, P. (1987) Campylobacter jejuni in dairy cows and raw milk. Epidemiol Infect 98(3):263-269. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 104 Humphrey, T.J. and Hart, R.J. (1988) Campylobacter and Salmonella contamination of unpasteurized cows' milk on sale to the public. J Appl Bacteriol. 65(6):463-467. Hussain, I., Mahmood, S.H., Akhtar, M. and Khan, A. (2007) Prevalence of Campylobacter species in meat, milk and other food commodities in Pakistan. Food Microbiology 24:219222. Husu, J.R. (1990) Epidemiological studies on the occurrence of Listeria monocytogenes in the feces of dairy cattle. Zentralbl.Veterinarmed.B 37(4):276-282. Hutchinson, D.N., Bolton, F.J., Hinchliffe, P.M., Dawkins, H.C., Horsley, S.D., Jessop, E.G., Robertshaw, P.A. and Counter, D.E. (1985) Evidence of udder excretion of Campylobacter jejuni as the cause of milk-borne campylobacter outbreak. J Hyg.(Lond) 94(2):205-215. Hutchison, M.L., Thomas, D.J., Moore, A., Jackson, D.R. and Ohnstad, I. (2005) An evaluation of raw milk microorganisms as markers of on-farm hygiene practices related to milking. J Food Prot 68(4):764-772. Ibrahim, A. and Macrae, I.C. (1991) Incidence of Aeromonas and Listeria-Spp in Red Meat and Milk Samples in Brisbane, Australia. International Journal of Food Microbiology 12(23):263-270. ICMSF (2002) Microorganisms in Foods 7: Microbiological Testing in Food Safety Management. In: International Commission on Microbiological Specifications for Foods. eds. Kluwer Academic/Plenum Publishers, New York. ICMSF. (1996) Microorganisms in Food 5: Microbiological Specifications of Food Pathgens. In: International Commission on Microbiological Specifications for Foods. eds. 1st ed, Blackie Academic and Professional, London. Jaros, P., Cogger, N. and French, N. (2008) A systematic review of the human disease evidence associated with the consumption of raw milk and raw milk cheeses. Massey University. Jayarao, B.M., Donaldson, S.C., Straley, B.A., Sawant, A.A., Hegde, N.V. and Brown, J.L. (2006) a survey of foodborne pathogens in bulk tank milk and raw milk consumption among farm families in Pennsylvania. Journal of Dairy Science 89(7):2451-2458. Jensen, N.E., Aarestrup, F.M., Jensen, J. and Wegener, H.C. (1996) Listeria monocytogenes in bovine mastitis. Possible implication for human health. International Journal of Food Microbiology 32(1/2):209-216. Jo, M.Y., Kim, J.H., Lim, J.H., Kang, M.Y., Koh, H.B., Park, Y.H., Yoon, D.Y., Chae, J.S., Eo, S.K. and Lee, J.H. (2004) Prevalence and characteristics of Escherichia coli O157 from major food animals in Korea. International Journal of Food Microbiology 95(1):41-49. Kabagambe, E.K., Wells, S.J., Garber, L.P., Salman, M.D., Wagner, B. and Fedorka-Cray, P.J. (2000) Risk factors for fecal shedding of Salmonella in 91 US dairy herds in 1996. Prev Vet Med 43(3):177-194. Kalman, M., Szollosi, E., Czermann, B., Zimanyi, M., Szekeres, S. and Kalman, M. (2000) Milkborne campylobacter infection in Hungary. J Food Prot. 63(10):1426-1429. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 105 Karns, J.S., Van Kessel, J.S., McClusky, B.J. and Perdue, M.L. (2007) Incidence of Escherichia coli O157:H7 and E.coli virulence factors in US bulk tank milk as determined by polymerase chain reaction. Journal of Dairy Science 90(7):3212-3219. Karns, J.S., Van Kessel, J.S., McCluskey, B.J. and Perdue, M.L. (2005) Prevalence of Salmonella enterica in bulk tank milk from US dairies as determined by polymerase chain reaction. Journal of Dairy Science 88(10):3475-3479. Keene, W.E., Hedberg, K., Herriott, D.E., Hancock, D.D., McKay, R.W., Barrett, T.J. and Fleming, D.W. (1997) A prolonged outbreak of Escherichia coli O157:H7 infections caused by commercially distributed raw milk. J Infect Dis 176(3):815-818. Khaitsa, M.L., Bauer, M.L., Lardy, G.P., Koetkott, D.K., Kegode, R.B. and Gibbs, P.S. (2006) Fecal shedding of Escherichia coli O157:H7 in North Dakota feedlot cattle in the fall and spring. Journal of Food Protection 69(5):1154-1158. Khan, A., Yamasaki, S., Sato, T., Ramamurthy, T., Pal, A., Datta, S., Chowdhury, N.R., Das, S.C., Sikdar, A., Tsukamoto, T., Bhattacharya, S.K., Takeda, Y. and Nair, G.B. (2002) Prevalence and genetic profiling of virulence determinants of non-O157 Shiga toxinproducing Escherichia coli isolated from cattle, beef, and humans, Calcutta, India. Emerg.Infect Dis 8(1):54-62. Klein, B.S., Vergeront, J.M., Blaser, M.J., Edmonds, P., Brenner, D.J., Janssen, D. and Davis, J.P. (1986) Campylobacter infection associated with raw milk. An outbreak of gastroenteritis due to Campylobacter jejuni and thermotolerant Campylobacter fetus subsp fetus. JAMA 255(3):361-364. Kongo, J.M., Malcata, F.X., Ho, A.J. and Wiedmann, M. (2006) Detection and characterisation of Listeria monocytogenes in Sao Jorge (Portugal) cheese production. Journal of Dairy Science 89:4456-4461. Korlath, J.A., Osterholm, M.T., Judy, L.A., Forfang, J.C. and Robinson, R.A. (1985) A pointsource outbreak of campylobacteriosis associated with consumption of raw milk. J Infect Dis 152(3):592-596. Kornblatt, A.N., Barrett, T., Morris, G.K. and Tosh, F.E. (1985) Epidemiologic and laboratory investigation of an outbreak of Campylobacter enteritis associated with raw milk. Am J Epidemiol 122(5):884-889. Lee, P.W. (1994) Teat disinfection usage by north eastern Victorian diary farmers. University of Melbourne. Lehner, A., Schneck, C., Feierl, G., Pless, P., Deutz, A., Brandl, E. and Wagner, M. (2000) Epidemiologic application of pulsed-field gel electrophoresis to an outbreak of Campylobacter jejuni in an Austrian youth centre. Epidemiol Infect 125(1):13-16. Lejeune, J.T., Hancock, D., Wasteson, Y., Skjerve, E. and Urdahl, A.M. (2006) comparison of E.coli O157 and shiga toxin encoding genes (stx) prevalence between Ohio, USA and norwegian dairy cattle. International Journal of Food Microbiology 109(1-2):19-24. Liewen, M.B. and Plautz, M.W. (1988) Occurrence of Listeria monocytogenes in raw milk in Nebraska. Journal of Food Protection 51(11):840-841. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 106 Losinger, W.C., Wells, S.J., Garber, L.P., Hurd, H.S. and Thomas, L.A. (1995) Management factors related to Salmonella shedding by dairy heifers. J Dairy Sci 78(11):2464-2472. Lovett, J., Francis, D.W. and Hunt, J.M. (1983) Isolation of Campylobacter jejuni from raw milk. Appl Environ Microbiol 46(2):459-462. Lovett, J., Francis, D.W. and Hunt, J.M. (1987) Listeria monocytogenes in raw milk; detection, incidence and pathogenicity. J Food Prot 50:188-192. Lund, A.M., Zottola, E.A. and Pusch, D.J. (1991) Comparison of methods for isolation of Listeria from raw milk. Journal of Food Protection 54(8):602-606. Magnusson, M., Christiansson, A., Svensson, B. and Kolstrup, C. (2006) Effect of different premilking manual teat-cleaning methods on bacterial spores in milk. Journal of Dairy Science 89:3866-3875. Makovec, J.A. and Ruegg, P.L. (2003) Results of milk samples submitted for microbiological examination in Wisconsin from 1994 to 2001. Journal of Dairy Science.2003 86(11):34663472. Marks, H.M., Coleman, M.E., Lin, C.T.J. and Roberts, T. (1998) Topics in microbial risk assessment:Dynamic flow tree process. Risk Analysis 18:309-328. Massa, S., Goffredo, E., Altieri, C. and Natola, K. (1999) Fate of Escherichia coli O157:H7 in unpasteurized milk stored at 8 degrees C. Lett Appl Microbiol 28(1):89-92. Maurin, M. and Raoult, D. (1999) Q fever. Clin.Microbiol.Rev. 12(4):518-553. Mazurek, J., Salehi, E., Propes, D., Holt, J., Bannerman, T., Nicholson, L.M., Bundesen, M., Duffy, R. and Moolenaar, R.L. (2004) A multistate outbreak of Salmonella enterica serotype typhimurium infection linked to raw milk consumption--Ohio, 2003. J Food Prot. 67(10):2165-2170. McEwen, S.A., Martin, S.W., Clarke, R.C., Tamblyn, S.E. and McDermott, J.J. (1988) The prevalence, incidence, geographical distribution, antimicrobial sensitivity patterns and plasmid profiles of milk filter Salmonella isolates from Ontaria dairy farms. Canadian Journal of Veterinary Research 52:18-22. McKee, R., Madden, R.H. and Gilmour, A. (2003) Occurrence of verocytotoxin-producing Escherichia coli in dairy and meat processing environments. Journal of Food Protection.2003 66(9):1576-1580. McManus, C. and Lanier, J.M. (1987) Salmonella, Campylobacter jejuni, and Yersinia enterocolitica in raw milk. J Food Prot. 50:51. Mechie, S.C., Chapman, P.A. and Siddons, C.A. (1997) A fifteen month study of Escherichia coli O157:H7 in a dairy herd. Epidemiol Infect 118(1):17-25. Meng, J., Doyle, M.P., Zhao, T. and Zhao, S. (2001) Food Microbiology: Fundamentals and Frontiers. Americian Society of Microbiology Press. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 107 Meng, J.S., Zhao, J.S., Doyle, M.P. and Joseph, S.W. (1998) Antibiotic resistance of Escherichia coli O157:H7 and O157:NM isolated from animals, food, and humans. J Food Prot 61:1511-1514. Meyer-Broseta, S., Diot, A., Bastian, S., Riviere, J. and Cerf, O. (2003) Estimation of low bacterial concentration: Listeria monocytogenes in raw milk. International Journal of Food Microbiology 80(1):1-15. MMWR. (1977) Outbreak of Milk-borne Salmonella gastroenteritis - South Australia. MMWR Weekly :127. MMWR. (1983) Campylobacteriosis Associated with Raw Milk Consumption -Pennsylvania . MMWR Weekly 32(26):337-338. MMWR. (1984a) Epidemiologic Notes and Reports Chronic Diarrhea Associated with Raw Milk Consumption -- Minnesota. MMWR Weekly 33(37);521-2,527-8(37):521-528. MMWR (1984b) Salmonella Dublin and raw milk consumption – California. MMWR Weekly 33(14):196-198. MMWR (2007a) Eschericha coli O157:H7 infection associated with drinking raw milk – Washington and Oregon, November – December 2005. MMWR Weekly 56(8):165-167. MMWR. (2007b) Salmonella Typhimurium infection associated with raw milk and cheese consumption – Pennsylvania, 2007. MMWR Weekly 56(44):1161-1164. MMWR (2008) Escherichia coli O157:H7 infections in children associated with raw milk and raw colostrum from cows – California, 2006. MMWR Weekly 57(23):625-628. Morgan, G., Chadwick, P., Lander, K.P. and Gill, K.P. (1985) Campylobacter jejuni mastitis in a cow: a zoonosis-related incident. Vet Rec. 116(4):111. Muraoka, W., Gay, C., Knowles, D. and Borucki, M. (2003) Prevalence of Listeria monocytogenes subtypes in bulk milk of the Pacific Northwest. Journal of Food Protection 66(8):1413-1419. Murinda, S.E., Nguyen, L.T., Ivey, S.J., Gillespie, B.E., Almeida, R.A., Draughon, F.A. and Oliver, S.P. (2002a) Molecular characterization of Salmonella spp. isolated from bulk tank milk and cull dairy cow fecal samples. J Food Prot. 65(7):1100-1105. Murinda, S.E., Nguyen, L.T., Ivey, S.J., Gillespie, B.E., Almeida, R.A., Draughon, F.A. and Oliver, S.P. (2002b) Prevalence and molecular characterization of Escherichia coli O157:H7 in bulk tank milk and fecal samples from cull cows: a 12-month survey of dairy farms in east Tennessee. J Food Prot. 65(5):752-759. Murinda, S.E., Nguyen, L.T. and Oliver, S.P. (2004) Problems in isolation of Campylobacter jejuni from frozen-stored raw milk and bovine fecal samples: genetic confirmation using multiplex PCR. Foodborne Pathog.Dis 1(3):166-171. Navratilova, P., Schlegelova, J., Sustackova, A., Napravnikova, E., Lukasova, J. and Klimova, E. (2004) Prevalence of Listeria monocytogenes in milk, meat and foodstuff of MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 108 animal origin and the phenotype of antibiotic resistance of isolated strains. Vet Med Czech 49(7):243-252. Nero, L.A., de Mattos, M.R., Beloti, V., Netto, D.P., Pinto, J.P.A.N., de Andrade, N.J., Silva, W.P. and Franco, B.D.G.M. (2004) Hazards in non-pasteurized milk on retail sale in Brazil: Prevalence of Salmonella spp, Listeria monocytogenes and chemical residues. Brazilian Journal of Microbiology 35(3):211-215. Newcombe, R.G. (1998) Two-sided confidence intervals for the single proportion: Comparison of seven methods. Statistics in Medicine 17:857-872. O'Donnell, E.T. (1995) The incidence of Salmonella and Listeria in raw milk from farm bulk tanks in England and Wales. Journal of the Society of Dairy Technology 18:25-29. Oksuz, O., Arici, M., Kurultay, S. and Gumus, T. (2004) Incidence of Escherichia coli O157 in raw milk and white pickled cheese manufactured from raw milk in Turkey. Food Control 15:453-456. Oosterom, J., Engels, G.B., Peters, R. and Pot, R. (1982) Campylobacter jejuni in cattle and raw milk in The Netherlands. Journal of Food Protection 45(13):1212-1213. Orr, K.E., Lightfoot, N.F., Sisson, P.R., Harkis, B.A., Tweddle, J.L., Boyd, P., Carroll, A., Jackson, C.J., Wareing, D.R. and Freeman, R. (1995) Direct milk excretion of Campylobacter jejuni in a dairy cow causing cases of human enteritis. Epidemiol Infect 114(1):15-24. Osterholm, M.T., MacDonald, K.L., White, K.E., Wells, J.G., Spika, J.S., Potter, M.E., Forfang, J.C., Sorenson, R.M., Milloy, P.T. and Blake, P.A. (1986) An outbreak of a newly recognized chronic diarrhea syndrome associated with raw milk consumption. JAMA 256(4):484-490. Padhye, N.V. and Doyle, M.P. (1991) Rapid procedure for detecting enterohemorrhagic Escherichia coli O157:H7 in food. Applied and Environmental Microbiology.1991 57(9):2693-2698. Peterson, M.C. (2003) Campylobacter jejuni enteritis associated with consumption of raw milk. J Environ Health 65(9):20-1, 24, 26. Pradel, N., Livrelli, V., De Champs, C., Palcoux, J.B., Reynaud, A., Scheutz, F., Sirot, J., Joly, B. and Forestier, C. (2000) Prevalence and characterization of Shiga toxin-producing Escherichia coli isolated from cattle, food, and children during a one-year prospective study in France. J Clin Microbiol 38(3):1023-1031. R Development Core Team (2007) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org. Rahn, K., Renwick, S.A., Johnson, R.P., Wilson, J.B., Clarke, R.C., Alves, D., McEwen, S., Lior, H. and Spika, J. (1997) Persistence of Escherichia coli O157:H7 in dairy cattle and the dairy farm environment. Epidemiol Infect 119(2):251-259. Renter, D.G., Morris, J.G., Sargeant, J.M., Hungerford, L.L., Berezowski, J., Ngo, T., Williams, K. and Acheson, D.W.K. (2005) Prevalence, risk factors, O serogroups, and MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 109 virulence profiles of Shiga toxin-producing bacteria from cattle production environments. Journal of Food Protection 68(8):1556-1565. Reuben, A., Treminio, H., Arias, M.L. and Chaves, C. (2003) Presence of Escherichia coli O157 : H7, Listeria monocytogenes and Salmonella spp. in food from animal origin in Costa Rica. Archivos Latinoamericanos de Nutricion 53(4):389-392. Rhoades, J.R., Duffy, G. and Koutsoumanis, K. (2009) Prevalence and concentration of verocytotoxigenic Escherichia coli, Salmonella enterica and Listeria monocytogenes in the beef production chain: A review. International-Journal-of-Food-Microbiology 26(4):357-376. Rice, D.H., Ebel, E.D., Hancock, D.D., Besser, T.E., Herriott, D.E. and Carpenter, L.V. (1997) Escherichia coli O157 in cull dairy cows on farm and at slaughter. Journal of Food Protection 60(11):1386-1387. Rodriguez, L.D., Fernandez Garayzabal, J.F., Boland, J.A.V., Rodriguez, F.E. and Fernandez, G.S. (1985) Isolation of micro-organisms of the species Listeria from raw milk intended for human consumption. Can.J Microbiol 31(10):938-941. Rohrbach, B.W., Draughon, F.A., Davidson, P.M. and Oliver, S.P. (1992) Prevalence of Listeria monocytogenes, Campylobacter jejuni, Yersinia enterocolitica, and Salmonella in bulk tank milk: risk factors and risk of human exposure. Journal of Food Protection.1992 55(2):93-97. Ross, T., Ratkowsky, D.A., Mellefont, L.A. and McMeekin, T.A. (2003) Modelling the effects of temperature, water activity, pH and lactic acid concentration on the growth rate of Escherichia coli. International-Journal-of-Food-Microbiology 82(1):33-43. Rugbjerg, H., Nielsen, E.M. and Andersen, J.S. (2003) Risk factors associated with faecal shedding of verocytotoxin-producing Escherichia coli O157 in eight known-infected Danish dairy herds. Prev Vet Med 58(3-4):101-113. Ryser, E.T. (2001) Public Health Concerns. In: Marth, E.H., Steele, J.L. eds. Applied Dairy Microbiology 2nd ed. Marcel Dekker, New York, pp 397-546. Salter, M.A. (1998) Effect of temperature and water activity on Escherichia coli in relation to beef carcasses. PhD Thesis, University of Tasmania, Australia. Sanaa, M., Poutrel, B., Menard, J.L. and Serieys, F. (1993) Risk factors associated with contamination of raw milk by Listeria monocytogenes in dairy farms. J Dairy Sci 76(10):2891-2898. Sato, K., Bartlett, P.C., Kaneene, J.B. and Downes, F.P. (2004) Comparison of prevalence and antimicrobial susceptibilities of Campylobacter spp. isolates from organic and conventional dairy herds in Wisconsin. Appl Environ Microbiol 70(3):1442-1447. Schlesser, J.E., Gerdes, R., Ravishankar, S., Madsen, K., Mowbray, J. and Teo, A.Y. (2006) Survival of a five-strain cocktail of Escherichia coli O157:H7 during the 60-day aging period of cheddar cheese made from unpasteurized milk. J.Food Prot. 69(5):990-998. Seglenloka, Z. and Dixon, S. (1977) Outbreak of Milk-borne Salmonella Gastroenteritis South Australia. Morbidity and Mortality Weekly Report :127. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 110 Shanson, D.C., Gazzard, B.G., Midgley, J., Dixey, J., Gibson, G.L., Stevenson, J., Finch, R.G. and Cheesbrough, J. (1983) Streptobacillus moniliformis isolated from blood in four cases of Haverhill fever. Lancet 2(8341):92-94. Sheldrake, R.F. and Hoare, R.J.T. (1980) Unknown. Journal Dairy Research 43(27). Shere, J.A., Bartlett, K.J. and Kaspar, C.W. (1998) Longitudinal study of Escherichia coli O157:H7 dissemination on four dairy farms in Wisconsin. Applied and Environmental Microbiology 64(4):1390-1399. Silva, I.M.M., Almeida, R.C.C., Alves, M.A.O. and Almeida, P.F. (2003) Occurrence of Listeria spp. in critical control points and the environment of Minas Frescal cheese processing. International Journal of Food Microbiology 81(3):241-248. Slade, P.J., Collins-Thompson, D.L. and Fletcher, F. (1988) Incidence of Listeria species in Ontario raw milk. Canadian Institute of Food Science and Technology Journal 21(4):425429. Small, R.G. and Sharp, J.C. (1979) A milk-borne outbreak due to Salmonella dublin. J Hyg.(Lond) 82(1):95-100. Sparks, N.H.C. (2009) The role of the water supply system in the infection and control of Campylobacter in chicken. Worlds Poultry Science Journal 65(3):459-473. Stanley, K., Cunningham, R. and Jones, K. (1998) Isolation of Campylobacter jejuni from groundwater. J Appl Microbiol 85(1):187-191. Strachan, N.J., Doyle, M.P., Kasuaga, E., Rotariu, O. and Ogden, I.D. (2005) Dose response modelling of Escherichia coli O157 incorporating data from foodborne and environmental outbreaks. International Journal of Food Microbiology 108:35-47. Tacket, C.O., Dominguez, L.B., Fisher, H.J. and Cohen, M.L. (1985) An outbreak of multiple-drug-resistant Salmonella enteritis from raw milk. JAMA 253(14):2058-2060. Tayganyilmaz., Moyer, B., Macdonell, R.E., Cordero-Coma, M. and Gallagher, M.J. (2009) Outbreaks associated with unpasteurized milk and soft cheese: An overview of consumer safety. Food Protection Trends 29(4):211-222. Taylor, D.N., Porter, B.W., Williams, C.A., Miller, H.G., Bopp, C.A. and Blake, P.A. (1982) Campylobacter enteritis: a large outbreak traced to commercial raw milk. West J Med 137(5):365-369. Thomas, C., Daughtry, B., Padula, D., Jordan, D., Arzey, G., Davey, K., Holds, G., Slade, J. and Pointon, A. (2006) Salmonella Quantitative Risk Assessment Model for the Australian Egg Industry. Australian Egg Corporation Limited, Sydney, Australia. http://www.aecl.org/images/File/Research%20Reports/SAR42A%20Final%20Report%20Foo d%20Safety%20Salmonella.pdf. Thurm, V., Stark, R., Maede, D., Fanghaehnel, S., Berger, W., Knobloch, H. and Lange, D. (2000) Raw milk as a source of foodborne Campylobacter infections. Bundesgesundheitsblatt-Gesundheitsforschung-Gesundheitsschutz. 43(10):777-780. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 111 Timm, C.D., Irino, K., Gomes, T.A.T., Vieira, M.M., Guth, B.E.C., Vaz, T.M.I., Moreira, C.N. and Aleixo, J.A.G. (2007) virulence markers and serotypes of shiga toxin producing Escherichia coli, isolated from cattle in Rio Grande do Sul, Brazil. Letters in Applied Microbiology 44:419-425. Uraz, G. and Yucel, N. (1999) The isolation of certain pathogen microorganisms from raw milk. Cent.Eur J Public Health 7(3):145-148. USDA/APHIS. (1994) Salmonella in Dairy Calves. USDA:APHIS, www.aphis.usda.gov/vs/ceah/cahm. USDA/APHIS. (2003a) Salmonella and campylobacter on US Dairy operations. USDA:APHIS, www.aphis.usda.gov/vs/ceah/cahm. USDA/APHIS. (2003b) Salmonella and listeria in bulk tank milk on US dairies. United States Department of Agriculture, Animal and Plant Health Inspection Service : USDA:APHIS, www.aphis.usda.gov/vs/ceah/cahm. Vaerewijck, M.J., De Vos, P., Lebbe, L., Scheldeman, P., Hoste, B. and Heyndrickx, M. (2001) Occurrence of Bacillus sporothermodurans and other aerobic spore-forming species in feed concentrate for dairy cattle. J Appl Microbiol 91(6):1074-1084. Van Donkersgoed, J., Berg, J., Potter, A., Hancock, D., Besser, T., Rice, D., LeJeune, J. and Klashinsky, S. (2001) Environmental sources and transmission of Escherichia coli O157 in feedlot cattle. Can.Vet J 42(9):714-720. van Kessel, J.S., Karns, J.S., Gorski, L., McCluskey, B.J. and Perdue, M.L. (2004) Prevalence of salmonellae, Listeria monocytogenes, and fecal coliforms in bulk tank milk on US dairies. Journal of Dairy Science 87(9):2822-2830. Vardar-Unlu, G., Unlu, M. and Bakici, M.Z. (1998) Incidence of Listeria spp. from raw milk in Sivas. Turkish Journal of Medical Science 28:389-392. Vissers, M., Driehuis, F., Te-Giffel, M.C., De Jong, P. and Lankveld, J. (2006) Improving farm management by modeling the contamination of farm tank milk with butyric acid bacteria. Journal of Dairy Science 89:850-858. Vissers, M.M., Driehuis, F., Te Giffel, M.C., De Jong, P. and Landveld, J.M.G. (2007) Quantification of the transmission of microorganisms to milk via dirt attached to the exterior of teats. Journal of Dairy Science 90:3579-3582. Vogt, R.L., Little, A.A., Patton, C.M., Barrett, T.J. and Orciari, L.A. (1984) Serotyping and serology studies of campylobacteriosis associated with consumption of raw milk. J Clin Microbiol 20(5):998-1000. Vose, D. (2008) Risk Analysis - A Quantitative Guide. In: Vose, D. eds. 3rd ed, John Wiley and Sons Pty Ltd, Wiltshire, England. Waak, E., Tham, W. and Danielsson-Tham, M.L. (2002) Prevalence and fingerprinting of Listeria monocytogenes strains isolated from raw whole milk in farm bulk tanks and in dairy plant receiving tanks. Applied and Environmental Microbiology 68(7):3366-3370. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 112 Wegmuller, B., Luthy, J. and Candrian, U. (1993) Direct polymerase chain reaction detection of Campylobacter jejuni and Campylobacter coli in raw milk and dairy products. Appl Environ Microbiol 59(7):2161-2165. Wells, J.G., Shipman, L.D., Greene, K.D., Sowers, E.G., Green, J.H., Cameron, D.N., Downes, F.P., Martin, M.L., Griffin, P.M. and Ostroff, S.M. (1991) Isolation of Escherichia coli serotype O157:H7 and other Shiga-like-toxin-producing E. coli from dairy cattle. J Clin Microbiol 29(5):985-989. Wesley, I.V., Wells, S.J., Harmon, K.M., Green, A., Schroeder-Tucker, L., Glover, M. and Siddique, I. (2000) Fecal shedding of Campylobacter and Arcobacter spp. in dairy cattle. Appl Environ Microbiol 66(5):1994-2000. Whyte, P., McGill, K., Cowley, D., Madden, R.H., Moran, L., Scates, P., Carroll, C., O'Leary, A., Fanning, S., Collins, J.D., McNamara, E., Moore, J.E. and Cormican, M. (2004) Occurrence of Campylobacter in retail foods in Ireland. Int J Food Microbiol 95(2):111-118. Wyatt, C.J. and Timm, E.M. (1982) Occurrence and survival of Campylobacter jejuni in milk and turkey. Journal of Food Protection 45(13):1218-1220. Xanthiakos, K., Simos, D., Angelidis, A.S., Nychas, G.J.E. and Koutsoumanis, K. (2006) Dynamic modeling of Listeria monocytogenes growth in pasteurised milk. Journal of Applied Microbiology 100:1289-1298. Yoshida, T., Kato, Y., Sato, M. and Hirai, K. (1998a) Sources and routes of contamination of raw milk with Listeria monocytogenes and its control. J Vet Med Sci 60(10):1165-1168. Yoshida, T., Sato, M. and Hirai, K. (1998b) Prevalence of Listeria species in raw milk from farm bulk tanks in Nagano prefecture. J Vet Med Sci 60(3):311-314. Zhao, T., Doyle, M.P., Shere, J. and Garber, L. (1995) Prevalence of Enterohemorrhagic Escherichia coli O157:H7 in a survey of dairy herds. Applied and Environmental Microbiology 61(4):1290-1293. Zhou, Z., Nishikawa, Y., Zhu, P., Hong, S., Hase, A., Cheasty, T., Smith, H.R., Zheng, M. and Haruki, K. (2002) Isolation and characterization of Shiga toxin-producing Escherichia coli O157:H7 from beef, pork and cattle fecal samples in Changchun, China. J Vet Med Sci 64(11):1041-1044. Zschock, M., Hamann, H.P., Kloppert, B. and Wolter, W. (2000) Shiga-toxin-producing Escherichia coli in faeces of healthy dairy cows, sheep and goats: prevalence and virulence properties. Lett Appl Microbiol 31(3):203-208. MICROBIOLOGICAL RISK ASSESSMENT OF RAW COW MILK 113